Date of submission:
University of Amsterdam
MSc Business Administration
“The effects of signals of quality on the
willingness-to-pay and purchase intentions”
2 | P a g e
Statement of Originality
This document is written by Student Bogdan Banica who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 | P a g e
The purpose of this study was to examine the indirect effect of different signals of quality, namely a quality label and a bio label, on the willingness to pay and purchase intentions via the perceived quality and perceived price. In order to investigate this, a between-group online experiment was designed in which four groups were created. The experiment consisted in showing the respondents photos on the same product in four different conditions in which the signals that appeared on the package were manipulated: no signal/standard package, presence of a quality sticker, presence of a bio label and presence of both signals, quality sticker and bio label. The results show that signals of quality influence the perception of quality and price, but they had no indirect effect on the purchase intentions. The quality label did not influence the perceived quality and perceived price, and, therefore, it had not influence on the willingness to pay. The bio label had a great impact on the perceived quality and perceived price and an indirect effect on the willingness to pay via these two variables. When both signals were used, the results were similar to the third condition when just the bio label was used. This indicates that the quality label has no effect on consumers' perceptions even if used in combination with other signals.
Keywords: willingness to pay (WTP), purchase intentions, willingness to buy, perceived price,
4 | P a g e
1. Introduction ... 6
2. Literature review ... 10
2.1 The concept of quality ... 10
2.2 Quality perception ... 12
2.3 Information asymmetry and signaling theory ... 14
2.4 Organic products ... 18 2.5 Willingness to pay ... 19 2.6 Perceived price ... 20 2.7 Purchase intentions ... 23 3. Conceptual model ... 25 3.1 Conceptual model ... 25 3.2 Hypotheses ... 26 4. Methodology ... 27
4.1 Purpose of the research ... 27
4.2 Research Design ... 27
4.3 Choice of product ... 30
4.4 Survey set up and measurement scales ... 30
4.5 Data collection ... 33
5. Results ... 34
5.1 Respondents’ characteristics ... 35
5.2 Relationship between signals of quality and perceived quality... 37
5.3 Willingness-to-pay ... 40
5.4 Perceived price ... 44
5.5 Purchase intentions ... 48
5.6 Hypotheses ... 50
6. Discussion and conclusion ... 51
6.1 Managerial contributions ... 54
6.2 Theoretical contributions ... 55
6.3 Limitations and further research ... 55
5 | P a g e
Annexes ... 70
Appendix 1 – Survey (Romanian version) ... 70
Appendix 2 – Survey (English version) ... 74
Appendix 3 – Quality sticker ... 78
Appendix 4 – Bio label... 78
Appendix 5 – Condition 1 ... 79
Appendix 6 – Condition 2 ... 80
Appendix 7 – Condition 3 ... 81
6 | P a g e
Willingness to pay and purchase intentions have been the main topic of discussion in many industries throughout the years. Researchers have identified several factors that can affect these two constructs, the most important one being the customer perceived value (Hsin Chang & Wang, 2011). The perceived value is mostly influenced by consumers’ perceptions of quality and price (Zeithaml, 1988). There is an information asymmetry between companies and consumers, which means that consumers are usually not aware of the quality of a product. Therefore, they do not observe the quality before purchase (Rao et al., 1999) unless it is communicated to them. One way of communicating the quality of a product is through signals of quality (Conelly, et al. 2011).
The use food labeling has become more and more important over the years and it allows companies to signal quality and charge a premium price. The reason for this is that the consumers have become increasingly concerned with their health and they demand safer and higher quality food products (Loureiro & McCluskey, 2000). According to the Council of the European Communities (Regulation 2081/92) consumers are more concerned with the quality of the food rather than the quantity. There are many signals that can influence consumers’ perception of quality. This study will focus on two signals: a bio label and an uncertified quality label. There are 2 types of quality labels, compulsory and voluntary. The compulsory ones are usually imposed by the government, while the voluntary ones are not obligatory for producers but they can create a competitive advantage (Grunert, 2005). Voluntary quality labels can be certified, which means they are “sponsored” by a third party such as a government organization or a private organization, and they can be uncertified, which means that no third party guarantees
7 | P a g e
for the quality. The bio label informs consumers that the product is “organic” which means that it was produced without any synthetic input.
The main concern of the researchers in the field of marketing was to provide entrepreneurs with a comprehensive overview of the possible ways to act and different combinations of the means of achieving the development of the company, taking into consideration their own resources and the external environment. This concern led to the birth of the “marketing-mix” concept, which currently holds a central position in the theory and practice of marketing. An important part of the marketing mix is the package. Therefore, the following research questions has been stated and investigated in this paper:
"How do different signals of quality, on the package, influence consumers' willingness to pay and purchase intentions for fast moving consumer goods?"
Furthermore, the following sub-questions were investigated:
Which signal is more efficient in increasing willingness to pay and willingness to buy? Do multiple signals used at the same time increase the perception of quality?
The setting for this research is the Romanian sugar market. In 1991, there were 33 sugar factories in Romania, but in 2015 the sugar market is divided between 4 main producers. In order to reach its EU quota, the production grew very rapidly between 2003 and 2007 by an average annual increase of 15%, amounting to 115 thousand tons in the latter year.
The main producers of sugar have their own sugar brand and they have their own packaging, but because their packaging capacity is not as high as their production capacity, they resort to wholesalers to pack and sell the remaining of the sugar. But these wholesalers sell sugar at lower
8 | P a g e
prices than the main producers because, even though it is the same product, the perceived quality of the producers’ brands is higher. The same principle applies to other fast moving consumer goods such as corn, rice etc. The customers view these wholesalers as producers because they package and distribute the products under their own brand. Therefore these intermediaries need a marketing strategy as well.
This paper investigates the impact of different signals on the consumers’ purchase intentions and willingness to pay for commodities, namely food products, in an experimental setting. Furthermore, this paper further explores the mechanism that determines consumers’ purchase behavior and willingness to pay. In order to do this, a between groups online experiment was designed. This experiment focused on the two signals previously mentioned, namely a quality label and a bio label. The effect of these signals was studied for a fast moving consumer good which was sugar. This effect is important to study because it seems that consumers make irrational purchase decisions that are based more on appearances than on the actual quality of the product inside the package. This phenomenon occurs because of the information asymmetry. Consumers are rarely able to detect quality at a glance and studies have shown that they engage in little information search, even when their financial commitment is substantial. Thus, the consumers rather rely on market quality signals such as advertising, brand popularity and price (Gerstner, 1985). Sugar is an excellent choice of product for this research because, on the Romanian market, there are many brands with many different prices, but the actual bulk product inside is the same. Before 2007 there might have been slight differences in the quality of these products, but all that is behind us since Romania entered the EU and all producers had to operate by certain standards imposed by the Union.
9 | P a g e
The study found that the purchase intentions for fast moving consumer goods were not influenced by signals of quality. Furthermore, the quality label had no effect on the willingness to pay, but the bio label had a major impact on the perceived quality and, therefore on the willingness to pay. When both signals were used at the same time, there was no significant difference in the perceived quality compared to when just the bio signal was used.
The main contributions of this study are the identification of socio-demographic characteristics that affect consumer preferences and it aims to analyze the efficiency of different signals in maximizing consumers’ willingness to pay and willingness to buy. From a managerial point of view it helps managers design better packages for their products in order to increase their customers’ willingness to pay and influence their purchase decision. From a theoretical point of view this paper adds to the literature by analyzing the mechanism that determines consumers to pay more for product and it investigates the relationship between the perceived price and perceived quality.
This thesis consists of 5 chapters and the introduction. The first chapter is the literature review in which the main constructs used in this research are explained and hypotheses are deducted from theory. The second chapter contains the conceptual model on which this research is based on and the main hypotheses that will be tested. The third chapter describes the methodology and research design. The fourth chapter is dedicated to the testing of the hypotheses using statistical analysis and the results are presented. The discussion is the last chapter of this paper and it contains the interpretation of the results, conclusions drawn and managerial and theoretical contributions. Furthermore this final chapter also contains the limitations of this study along with recommendations for future research.
10 | P a g e
2. Literature review
This chapter starts with an introduction to the concept of quality. Afterwards a distinction between actual quality and the perception of quality is made. The third subchapter explains the signaling theory and how it can be used to reduce information asymmetry. The fourth subchapter summarizes the concept of organic products and how they are obtained followed by a discussion about willingness to pay. The last two chapters are dedicated to the perceived price and purchase intentions which are two of the dependent variables focused on in this study. This literature review provides arguments for the hypothesis that were tested in this research.
2.1 The concept of quality
As Lancaster (Lancaster, 1971) portrayed, product quality is a heap of characteristics that determine products' performance. At a small scale, product quality has been viewed as a basic variable for producers and customers (Steenkamp, 1990). These days, managers depend on their products' remarkable quality to undermine their competitors. Regardless of the industry, size of the organization and type of production, quality has turned into a vital component of the competitive strategy (Wolff, 1986). According to Porter (2008), superior quality is an efficient technique to create a competitive advantage, gain the loyalty of the customers and create entry barriers for potential competitors. Furthermore, Peter and Waterman (1982) distinguish quality as one of the main variables that can determine the success or the failure of a company.
Research has shown that there are three types of attributes that characterize the quality of a product: search, experience and credence (Nelson, 1970; Nelson, 1974). Search attributes are the ones that can be inspected prior to the purchase like the package design or the shape of the product. Attributes that can be distinguished post-purchase and post-consumption, such as taste
11 | P a g e
that cannot be evaluated before consumption, are experience attributes. Regardless of the type of product, consumers are not able to determine the credence attributes not even after consumption (Dimara & Skuras, 2005). An example of such an attribute is whether the meat is healthy and safe to eat. This cannot be checked by regular consumers, but instead they have to rely on the evaluation of other competent parties (Grunert, Bech-Larsen, & Bredahl, 2000). Other researchers refer to the search, experience and credence attributes of quality as intrinsic and extrinsic quality cues. The differentiate between these two, so that the intrinsic cues refer to physical attributes of the product that cannot be changed unless the product itself is changed such as technical specifications, taste or color (J. C. Olson, 1977), whereas the extrinsic cues refer to all the other characteristics of a product such as price, brand etc. which are determined by marketing efforts (Holm & Kildevang, 1996; Steenkamp, 1989).
Food quality has been the main focus of many researchers over the years. Studies have shown that the two aspects of food quality, experienced and expected, have four main dimensions (Grunert, Larsen, Madsen, & Baadsgaard, 1996): appearance and taste, process, health and convenience. Firstly, the attributes of taste and appearance can be characterized as hedonic qualities. This dimension is viewed, for the most part, as an experienced attribute of food products, on the grounds that taste can be perceived strictly after consumption and despite the fact that the food's appearance can be perceived before eating. Secondly, more and more consumers see the production process as an increasingly important quality dimension. This dimension looks at the welfare of animals and at the organic production process. Clearly, this is a credence attribute because of the information asymmetry which means that consumers usually know nothing about the production process and they have to trust other parties to convey this information (Brunsø, Fjord, & Grunert, 2002). Furthermore, in the recent years, the health
12 | P a g e
dimension has become as important as the taste and appearance and it includes both functional characteristics and safety and risk related issues. This dimension includes credence attributes, since the consumers cannot usually asses by themselves the impact of food products on their health, not even after consumption (Brunsø et al., 2002). Various studies have shown that consumers are inclined to take into consideration this dimension when buying food, propelled by desires of both higher quality and a longer life (Roininen et al., 2001). Finally, convenience is another progressively important dimension of food quality and can be characterized as saving of time, mental or physical energy. Thus, it is very difficult to achieve convenience and it is credence attribute as well (Gofton, 1995).
2.2 Quality perception
In general, quality can be defined as superiority or excellence. The perceived quality is different from objective or actual quality. It can be defined as the consumer's judgment about a product's overall excellence or superiority, while "objective quality” represents the actual excellence of the products or technical superiority (Zeithaml, 1988). Other researchers define perceived product quality "as the way in which a customer views a product's brand equity and overall superiority compared to the available alternatives"(Beneke, Flynn, Greig, & Mukaiwa, 2013).
According to Cardello (1995), food quality is strongly connected to the concept of acceptability can also be termed as quality perception. The perceived quality is commonly seen as an attitude (Ophuis & Van Trijp, 1995). According to Gotlieb et al. (1994), "perceived quality is formed through a process similar to one in which attitudes are formed". Thus, the assessment of perceived quality can be affected by the characteristics that are associated with a product and the appraisal of these characteristics (Dodds, Monroe, & Grewal, 1991).
13 | P a g e
According to Oude Ophuis and Van Trijp (1995) the perceived quality is influenced by four modalities which are represented in the following figure:
Figure 1: Quality Quadrant Source: ( Ophuis & Van Trijp, 1995, p. 178)
Firstly, the perceived quality is influenced by the Perception process. The general evaluation is based on the product attributes that were experienced or are thought to be associated with the evaluated product and they that can be visible or invisible. Secondly, the perception of quality varies depending on the type of Product. For example, when looking at food products, the content of fat can be seen as a quality characteristic for meat, but it is not applicable for fruits and vegetables. Moreover, the perceived quality is different depending on the Person that evaluates it because consumers have different perceptions based on personal preferences and experience. To give an example, a wine connoisseur will have different perceptions of wine in comparison to a regular consumer. Lastly, perceived quality can be influenced by circumstances which were summarized in the previous figure as Place. The intended purpose for the product accounts for a variation in the perception. For instance, overripe tomatoes can be considered
14 | P a g e
good quality for soups and sauces, but they are not appropriate for fresh salads (Ophuis & Van Trijp, 1995).
The concept of quality indicators is a main component when talking about perceived quality. Consumers cannot asses all the dimension on which the concept of quality is based on. Therefore, in order to judge the perceived product quality, consumers use indirect or substitute indicators of quality (Ophuis & Van Trijp, 1995). According to Olson (1972) the perception of quality process takes place in two stages. First of all, the consumers choose substitute indicators of quality (e.g. quality cues) from a variety of attributes related to the product, after which they combine the assessment of these cues to form a general evaluation of the product quality. Later, Steenkamp (Steenkamp, 1989) developed Olsen's work further and differentiated between quality cues and quality attributes, so that the former can be evaluated before consumption, while the latter can only be observed after consumption. Moreover, he considers quality attributes perceptions to be the basis on which the overall quality judgments are formed.
2.3 Information asymmetry and signaling theory
The information asymmetry between two parties (individuals or organizations) influences the way they behave and in order to best describe this behavior, the signaling theory is useful (Connelly, et al., 2011). This theory is basically concerned with reducing the information asymmetry between the parties (Spence, 2002), but management scholars applied it to help explain different effects of information asymmetry in a wide range of research contexts (Connelly, et al., 2011).
The decision-making process that is used by individuals, businesses or governments is influenced by the information they have available. As it was explained by Stiglitz (2002), the
15 | P a g e
information asymmetry arises when “different people know different things”. Furthermore, he emphasizes two main types of information where the asymmetry is most important: information about quality and information about intent (Stiglitz, 2000). This paper will focus on the information about quality and how signaling this information can affect the consumers’ willingness to pay.
In most cases the quality of a product is not observable to customers before the purchase, but it is uncovered fully after the purchase (Rao, et al., 1999). Each company has the choice of whether to signal the true quality of their products or not. While it may be beneficial for high-quality firms to signal their quality, it could potentially harm a lower-quality firm (Connelly, et al., 2011).
The signaling theory has three main elements: the signaler which is an insider (e.g. executives or managers) that has information that is not available to everybody and decides to share it; the
signal which is the action of conveying the information to outsiders; the receiver who is an
outsider that lacks the information but would like to receive it (Connelly, et al., 2011). The signalers and receivers can have, to a certain extent, conflicting interests because the successful misleading of the receiver could benefit the signaler (Bird, et al., 2005). Thus, inferior signalers are motivated to act dishonestly intentionally by giving fake signals so that the receivers will select them (Johnstone & Grafen, 1993). On the other hand, the signal is usually credible and informative, because those that are caught trying to signal dishonestly will face unfavorable monetary consequences.
Management scholars suggest that signals of quality can be "strong" or "weak", depending on how easily they can be detected by the receiver compared to other signals (Gulati & Higgins,
16 | P a g e
2003). Other researchers make a distinction between signal strength and visibility. The signal
strength is described as how important the signal is for a certain signaler, while the visibility
refers to how easily can the signal be spotted or observed (Ramaswami, et al., 2010).
On the other hand, the signaling effectiveness is also influenced by the receiver's characteristics. In order for the signaling process to work, the receiver has to look for the signal and he has to know what to look for (Connelly, et al., 2011). For weak signals which can be difficult to observe if the receiver is not looking for them, monitoring the environment is especially important (Ilmola & Kuusi, 2006), but once a signal is received and used in order to make an informed choice, it is more likely that similar signals will be spotted in the future (Cohen & Dean, 2005). Other researchers observed that different receivers interpret signals differently (Perkins & Hendry, 2005).
Under information asymmetry, brand names can act as a signal of quality that cannot be observed (Rao, et al., 1999). In marketing studies, the receivers are the consumers (Connelly, et al., 2011). Consumers can punish a brand if they associate it with high quality products and it turns out that the products are of a lower quality than expected (Montgomery & Wernerfelt, 1992). These punishments can vary from bad word of mouth to abstaining from future purchases of that particular brand (Wernerfelt, 1988). Brand names are predominantly good long-term marketing investments and they can serve as an important marketing tool (Kotler, 2000). Furthermore, because these punishments can have detrimental effects on the revenues of the seller, brand names can also act as reinsurance or a signal of a product’s unobservable quality. Thus, branded products are more likely to be of a higher quality that the unbranded ones (Rao, et al., 1999). The more money was spent to build a brand’s reputation, the more credible the signal of quality is (Ippolito, 1990).
17 | P a g e
Another important aspect that is taken into consideration, when consumers judge and decide how much a product is worth, is the seller. That implies that the assessment of the value of the product is also based on secondary product attributes and the associations held towards a high status store extend to the sold goods (Silje & Havard, 2012). There are several empirical studies that have proven that the buyers’ perceived quality of a product is higher when the perceptions of store names are more favorable (Chu & Chu, 1994). When a producer doesn’t have a recognized brand name, this effect of the retailer on the perceived quality of a product is even stronger (Dodds, et al., 1991). Thus, the reputation of a retailer can be “rented” by a manufacturer in order to signal product quality (Chu & Chu, 1994).
There are a wide range of other signals of quality that the management scholars have identified. According to Certo (2003), companies must strive for legitimacy in order to survive. One way of achieving this legitimacy is signaling unobservable quality through a prestigious board of directors (Certo, et al., 2001) or prestigious top managers (Lester, et al., 2006). Furthermore, another signal of quality can be positive reputation of the company (Coff, 2002). Other common signals of quality can be a firm's owners (Sanders & Boivie, 2004) or in the case of start-ups the founder ownership (Busenitz, et al., 2005), management stability (Perkins & Hendry, 2005), intellectual property (Warner, et al., 2006) and inter-organizational ties (Gulati & Higgins, 2003; Park & Mezias, 2005).
As it was stated before, in most cases the quality of a product is not observable to customers before the purchase (Rao, et al. 1999). Therefore, signals of quality are important because they reduce the information asymmetry between companies and consumers and influence the perception of quality. Furthermore, by increasing the signal frequency, which means sending
18 | P a g e
more signals, the signaling effectiveness can be enhanced (Janney & Folta, 2003). Based upon this literature review, the following hypotheses are formulated:
H1a: The presence of a signal of quality has a positive effect on the perceived quality H1b: The quantity of signals of quality has a positive effect on the perceived quality
2.4 Organic products
Another signal of quality could be considered an “organic label” (bio-product). Even though this is not the main purpose of such a label, it is usually seen as a signal of quality by consumers. ”Organic” refers to food products that are produced without any synthetic input. Because only natural inputs are allowed the production is diminished and more work is required (Food and Agriculture Organization, 1999).
In most cases, the organic products are identical with the regular ones and consumers are not able to detect the organic characteristics not even after consumption. In other words, consumers only know whether a product is organic or not if they are told so (Giannakas, 2002). The demand for organic products has increased rapidly since mid-1989 because the health and environmental problems have become popular in the EU and the United States (Kalogeras, Valchovska, Baourakis, & Kalaitzis, 2009), but the high price of such products is considered the main inhibitor of organic consumption (Schifferstein & Ophuis, 1998). Other general barriers against consumption can be low availability and poor appearance (Zanoli & Naspetti, 2002).
The main reasons of the consumers that buy and consume organic products are health related and environmental concerns (Tregear, Dent, & McGregor, 1994). Other studies have shown that American consumers, who believe that conventional food poses a health risk, are willing to pay
19 | P a g e
more for organic food without pesticides (Jolly, 1991). Therefore, the following hypotheses are formulated:
H2a: The relationship between signals of quality and perceived quality is moderated by the type of signal.
H2b: In the absence of any prior knowledge about the brand, the organic label has a higher positive effect on the perceived quality than other signals of quality
2.5 Willingness to pay
There are several factors that influence consumer's willingness to pay (WTP), most common being the price that has to be paid for a product. Zeithaml (1988) stated that from a consumer's point of view, the price is what is sacrificed or given up in order to obtain a product.
According to Krishna (1991a), the willingness to pay is the maximum amount of money consumers are willing to spend for a product or service, but other researchers argue that there are different types of willingness to pay. They distinguish between "actual" WTP and hypothetical WTP. Hence, the maximum amount of money consumers would pay in a real shopping experience is the actual WTP, while the hypothetical one which is most used in market research refers to consumers' "statement" of how much they are willing to pay. This second type is less reliable because consumers have the knowledge that their statement will have no economic consequences (Miller & Hofstetter, 2009).
Most of the research on the competition between manufacturers and private labels in the retail sector focused on product quality as the answer for fast moving consumer goods that want to avoid a price war (Ghose & Lowengart, 2001; Steenkamp, Van Heerde, & Geyskens, 2010). The
20 | P a g e
main idea is that increasing the perceived product quality, a brand can create a competitive advantage that motivates consumers to pay more. Furthermore, in order to build strong brands within the food industry, the brand managers focus on quality image (Anselmsson & Lars Anders, 2013). However, in many cases, the manufacturing brands are producing for the competing private label and that means that the only difference between them is the brand name (Anselmsson, Vestman Bondesson, & Johansson, 2014).
Thus, quality is not such an efficient competitive tool anymore (Gerzema & Lebar, 2008) and it is not enough to sustain a competitive advantage. Perceived quality can only explain a fraction of the price premium the consumers are willing to pay for different fast moving consumer goods (Sethuraman, 2003). Other studies have also shown that the difference in price between manufacturer brands and store brands cannot be explained only by the perceived quality (Anselmsson et al., 2014).
Anselmsson et al. (2007) suggest that consumers' willingness to pay for food brands is influenced by five factors: perceived quality, awareness, uniqueness, loyalty and non-product related brand associations to corporate social responsibility. Taking this into consideration, the following hypothesis is formulated:
H3: There is a positive relationship between the perceived quality and the willingness-to-pay
2.6 Perceived price
From the consumers' point of view, pricing researchers defined price as a sacrifice. The price is composed of multiple elements such as objective price, perceived nonmonetary and monetary price and sacrifice as it is depicted in Figure 2 (Zeithaml, 1988). The objective price which is the
21 | P a g e
actual price of a good is different than the perceived price which is how consumers view the price and can vary from one consumer to another (Jacoby & Olson, 1977).
Figure 2: A Means-End Model Relating Price, Quality and Value Source: (Zeithaml, 1988, p. 4)
As they do not always remember the actual price of a good, consumers record prices in a way that has meaning to them (Dickson & Sawyer, 1986). The importance of price depends on the product category and demographic groups. Studies have shown that the price is not so important when buying fast moving consumer goods (Dickson & Sawyer, 1986), but the price awareness is greater when the consumers are older married females that do not work outside their home (Zeithaml & Berry, 1987). Another reason for the inconsistency between the objective price and the perception of price is the fact that the same brands or similar products have different prices across different stores (Maynes & Assum, 1982).
22 | P a g e
Research has shown that a higher price can negatively affect the willingness to buy because it represents a financial burden (Völckner & Hofmann, 2007) , but at the same time it is used by consumers as a quality cue (Völckner & Sattler, 2005). The relationship between price and perceived quality is lower for services than for fast moving consumer goods (Bijmolt, Heerde, & Pieters, 2005). This can be explained by the fact that the consumers are not so motivated to engage in a profound decision-making process when it comes to fast moving consumer goods, but instead they rely on cues that they can easily recognize such as price. This way the shopping experience becomes more convenient because of the simplified process, but as the consumers become more familiar with the product, this effect declines. Therefore, lowering the price not only lowers consumer costs, but it can also cause them to doubt the quality of the product by making negative price-quality inferences (Völckner & Hofmann, 2007). Many studies focused on the relationship between the price and the perceived quality, but little research was done regarding the relationship between the perceived price and the perceived quality. Despite the common belief that the price is an indicator of quality, many of the studies about the relationship between the price and the perceived quality had contradicting results and no general relationship between them was found (Verma & Gupta, 2004). Thus, the following hypothesis is formulated:
H4: There is a positive relationship between the perceived quality and the perceived price
The perception of price is influenced by numerous factors including determinants of perceived value and reference prices used to assess given prices. The willingness to pay for a product is mainly determined by a subjective value of the preferred option because consumers usually have a variety of choices (Simonson & Drolet, 2004). Furthermore, researchers found evidence that price sensitivity is considerably influenced by consumers’ price perception (Munnukka, 2008). Other researchers defined the perceived price as the internal reference price (Monroe, 1973),
23 | P a g e
which was found to have a great impact on the willingness to pay (Ranyard, Charlton, & Williamson, 2001). Therefore, the following hypothesis is formulated:
H5: There is a positive relationship between the perceived price and the willingness-to-pay
2.7 Purchase intentions
According to Assael (1995), the purchase intentions represent plans to buy a product or a service in the future and can be used to forecast which brand a consumer will purchase. They are influenced by beliefs and attitudes towards particular products (Fishbein & Ajzen, 1975). Attitudes are formed as a consequence of affective, cognitive and behavioral elements. Consumers' purchase behavior is influenced by affective and cognitive factors which mean that the purchase decisions are based on both emotions and rationality (Li, Monroe, & Chan, 1994). The purchase intentions will also be referred as willingness to buy (WTB) in this paper.
Gardener (1971) analyzed, in an experimental setting, the relationship between the price and quality for three products: a man's shirt, toothpaste and a suit. While he did not find any influence of price over the perceived quality, he found a relation between price and willingness to buy a shirt. Furthermore, it was found that in the purchase-decision process, the price has more than one role. According to the traditional economic theory, a higher price has a negative effect on buying decision because it affects consumers' budget and it is seen as a sacrifice, but it can also be perceived as an indicator of quality (Monroe & Krishnan, 1985). Therefore, price can have a negative or a positive effect on the purchase decision (Rao & Monroe, 1988).
Previous research suggests that the perceived price is characterized as an extrinsic cue and it is one of the most important types of information that is available to consumers during the purchase-decision process (Ralston, 2003; Wangenheim & Bayón, 2007). Zeithaml (1988)
24 | P a g e
emphasizes that the price seen as a sacrifice is the most significant factor when consumers assess the perceived value of a product. Furthermore, numerous other researchers found a strong negative relationship between the perceived price and the perceived value of a product because the purchasing power is diminished by a higher price (Boksberger & Melsen, 2011; Desarbo, Jedidi, & Sinha, 2001; Kashyap & Bojanic, 2000). The price is an indication of how much consumers need to pay for a product or servie. Tam (2004) also found that there is a negative relationship between costs and perceived value. The perceived value was defined by several researchers, such as Zeithaml (1988) who stated that value is “the consumer’s overall assessment of the utility of a product based on perceptions of what is received and what is given”. Another conceptualization of customer value belongs to Woodruff (1997) who defined it as “a trade-off between of benefit, the received component, and sacrifices, the given component”.
Customer perceived value is possibly the most important factor that influences the willingness to buy (Hsin Chang & Wang, 2011). Dodds and Monroe (1985) suggest that there is a positive relation between perceived value and willingness to buy. Beneke et al. (2013) also found that the perceived value has a significantly positive effect on the willingness to buy.
Erickson and Johansson (1985) explored the effect of price on the purchase intentions and found there is a positive indirect relationship between them because the price positively influences the perceived quality which in turn has a positive effect on attitude which determines the willingness to buy. Furthermore, they found that because of the budget constraints, the perceived price has an independent and negative effect on the willingness to buy a car. Therefore, the following hypothesis is formulated:
25 | P a g e
3. Conceptual model
Based on the literature review above, a conceptual model (Figure 3) has been created that summarizes the relationships hypothesized in this paper. As indicated in the conceptual model the signals of quality, have an indirect effect on willingness to pay and on willingness to buy via the two dependent variables perceived quality and perceived price. The relationship between signals of quality and perceived quality is moderated by the type of signal. There is a direct effect of the perceived quality on the willingness to pay, but this relationship is also mediated by the perceived price. Furthermore, it is hypothesized that the relationship between the perceived quality and willingness to buy is mediated by the perceived price.
3.1 Conceptual model
26 | P a g e
H1a: The presence of a signal of quality has a positive effect on the perceived quality
H1b: There is a positive relationship between the quantity of signals of quality and the perceived quality
H2a: The relationship between signals of quality and perceived quality is moderated by the type of signal.
H2b: In the absence of any prior knowledge about the brand, the organic label has a higher positive effect on the perceived quality than other signals of quality
H3: There is a positive relationship between the perceived quality and the willingness-to-pay H4: There is a positive relationship between the perceived quality and the perceived price H5: There is a positive relationship between the perceived price and the willingness-to-pay H6: There is a negative relationship between the perceived price and the willingness-to-buy
27 | P a g e
This chapter starts with the purpose of the research followed by the research design where the experiment that was conducted for this study is explained. The next subchapter discusses the choice of product for the experiment followed by a discussion about the survey and how the variables used in the research were measured. Finally, the method used to collect the data is discussed in the last subchapter.
4.1 Purpose of the research
This research aims to analyze the effects signals of quality have on the willingness to pay and purchase intentions. Furthermore, the influence of consumers' characteristics on these two variables was analyzed. In this research, the bio label which indicates that the product is organic was used as a signal of quality. In theory, a higher number of signals of quality will increase the perceived quality and therefore the willingness to pay and willingness to buy. This study clarifies whether signals of quality on the package are an efficient marketing tool in the fast moving consumers goods sector that can create a competitive advantage in an over saturated market. Moreover, this investigation helps to better understand the mechanism that determines consumers to pay more for the same product taking into consideration only the design of the package and the information conveyed to the consumers.
4.2 Research Design
In order to test the effects of signals of quality on willingness to pay and purchase intentions, a quantitative research method was adopted. These effects were explored in an experimental setting that was designed after a research question was stated. According to Field (2009), the beginning of a research consists in making observations and formulating a research question.
28 | P a g e
Based on a primary literature review and observations made in supermarkets, the following research question was formulated:
"How do signals of quality, on the package, influence consumers' willingness to pay and purchase intentions for fast moving consumer goods?"
The next step after stating the research question was an extensive literature review that explained the research question. Based on it eight hypotheses were formulated and tested by conducting an online experiment. Therefore, this is a deductive research because hypotheses were deducted after studying previous research (Saunders, Saunders, Lewis, & Thornhill, 2011). An experiment was chosen for this research because it made possible to expose different consumers to a similar product with small differences in the package design and compare their responses.
The advantages of an online experiment include lower costs, less time needed to set up the experiment and most important is that it is very easy to distribute the survey and collect responses in a short time. There are also disadvantages to this approach, especially when measuring the perceived quality since the intrinsic cues are excluded from the experiment. Furthermore, there is a big chance that respondents will not fill in the entire survey because they are not interested or motivated enough. This effect was alleviated in this experiment by making the survey relatively short. The average time of completion was five minutes.
In order to test the hypotheses two different signals of quality were used: a signal that clearly suggests that the product is of a higher quality (Appendix 3) and an organic label (Appendix 4). Different respondents were used in each treatment. This means that each treatment of the experiment includes a different group of participants. This was done by random allocation, which ensured that each participant had an equal chance of being assigned to any of the groups.
29 | P a g e
A between-group experimental design's main advantage is that different variables can be tested simultaneously, while the main disadvantage is that in order to get reliable results it is required to have at least 30 respondents for each treatment (Saunders et al., 2011).
The participants of this experiment were exposed to four different treatments (only one treatment per participant). The survey was the same for all of them except for the fourth page where the package that was presented was different. In the first condition where the respondents were presented with the plain sugar package (Appendix 5) with no additional signal of quality is the baseline for the experiment (control group). For the following treatments, the independent variable that was manipulated is the quality signal on the package (type of signal and number of signals) as it is presented in the following table:
Table 1: Overview of the experimental design
variablesPlain sugar package with no signal of quality (Appendix 5) Sugar package with one signal of
quality (Appendix 6)
Sugar package with an organic label (bio product)
Sugar package with two signals of quality (organic
label + high quality sticker)
variables Willingness to pay Perceived price Perceived quality Willingness to buy
30 | P a g e
4.3 Choice of product
The product that was chosen for this experiment was sugar. In order to eliminate the effect of the brand or product familiarity, a new package was designed with a non-existent brand (Appendix 5). Therefore, previous knowledge about the brand did not influence participants’ responses. Sugar is a good choice for this experiment because in the sugar industry there isn’t much advertising and usually the main factor that influences the decision to buy and the perceived quality is the design of the package and the information displayed on it. This product is commercialized under many different brands with many different prices, but the bulk product inside the package is essentially the same. This makes sugar one of the best choices for this research because the intrinsic cues do not have a role in consumers’ purchase-decision process or in deciding how much they are willing to pay for it. Therefore, the internal validity of the experiment is increased.
4.4 Survey set up and measurement scales
The survey contains several types of questions: multiple choice questions (single answer and multiple answers), 10 point Likert scale questions and open questions. The first page contains a statement from the researcher where the respondents are informed about the research they are going to participate in and they are assured that all the data that is collected is anonymous and will not be used for other purposes. The second page contains demographic questions about age, sex, income, education, occupation, marital status and how many people live in their household. Furthermore, this page contained two questions about how much sugar they consume, on average, per month and what type of sugar they usually use (organic or regular sugar). On the third page of the survey participants are shown a picture of the product followed by questions used to measure the dependent variables. The third page ends with a manipulation check that
31 | P a g e
consists in a multiple choice question (multiple answers are possible) where participants are asked to select what signals they noticed on the package they were shown at the beginning of the page. The questions were formulated in Romanian since the survey was distributed in Romania and it was pilot tested with representatives of each age category to make sure all of them are understandable to everyone. The survey was also translated into English. Both versions of the survey, the Romanian and English translations, can be found in the Appendix section of this paper (Appendix 1; Appendix 2).
In order to measure the willingness to pay and the perceived price the stated preference method was used. This method can also be defined as a direct survey. The data collected this way is often called revealed preference data (Breidert, Hahsler, & Reutterer, 2006). Another method of measuring the willingness to pay is the vickrey auction in which the highest bidder wins the auction, but the actual price paid at the end is the second bid. The first method (direct survey) was chosen for this research because the experiment is based on a product that does not exist. The respondents were asked to base their evaluation of the product based solely on the package and express their WTP and perceived price in the Romanian national currency which is RON. The exchange rate between RON and Euro is the following: 1 Euro= 4.4 RON. The questions that were used to measure these two constructs are the following:
How much would you be willing to pay for this product?
How much do you think this product costs at a retailer/supermarket?
Based to the literature review, the perceived quality is determined by intrinsic and extrinsic cues. Extrinsic cues are non-product related characteristics whereas intrinsic cues are product related. In this research the extrinsic cues were manipulated in order to influence the perception of quality. The extrinsic cue that was manipulated is the design of the package, namely the signals
32 | P a g e
of quality that appear on it. In order to measure this construct 7 items which consist of statement based questions on a 10 point Likert scale were used with a Cronbach's Alpha of 0.947. The respondents were one again asked to answer these questions basing their evaluation of the product solely on the package and compare it with the type of sugar they usually use. The following items were used in the survey:
This product is different……. This product has a better taste……. This product has a higher price……. This product is healthier…….
With this product you will make tastier cookies……. With this product you will make a better coffee……. The overall perceived quality of this product is…….
The willingness to buy or purchase intentions represent future plans to buy a certain product and it is taken as an attitude. In previous studies this construct was usually measured by questions on a 10 point Likert scale. In this survey it was measured with a single item on a 10 point Likert scale:
How likely is it that you will buy this product at a supermarket/retailer for personal consumption?
33 | P a g e
4.5 Data collection
The technique used in this experiment is the probability sampling which means that every respondent had an equal chance of being assigned to any of the conditions. This technique is frequently associated with experimental and survey research strategies (Saunders et. al, 2011). The experiment consists in an online survey that was distributed in two ways. The first method of distribution was through social media and e-mails and the second one consisted in going into supermarkets with an Ipad connected to the internet and ask consumers who added sugar in their shopping cart to fill in the survey on the spot. The online tool Qualtrics was used to design the survey and distribute it to the target sample. The Qualtrics software was used for quantitative statistical analysis in many professional and academic journals and it enables all kinds of online data collection and analysis including market research. After deleting the uncompleted questionnaires, the total number of respondents was 229 with 55 being assigned to the first condition, 57 to the second condition, 58 to the third condition and 59 to the fourth condition. The data collected for this experiment is primary data.
34 | P a g e
This chapter is going to describe the steps that were taken to analyze the data for this study and present the results of this analysis. The purpose of this research is to examine the influence of the independent variable “signals of quality” on the dependent variables “perceived quality”, “perceived price”, “willingness to pay” and “willingness to buy.
The first step in analyzing the data was to do a frequency check for all the variables in order to see if there are any errors in the data and if there any missing values. There were some errors in the data and some missing data. The most common error was that the participants filled the answer box of willingness to pay and perceived price with text to denominate the currency they used when filling the survey. Furthermore, they used text to denominate what unit of measure they used when they answered the question about how much sugar they consume on average per month. In total there were 56 errors that were corrected manually. In order to deal with the missing data, 11 cases were discarded because they were missing too many values. Moreover, the remaining amount of missing data was less than 10% for all variables. Therefore, Hotdeck imputation was used to complete the dataset.
The next step was to do a reliability analysis for the items that were used to measure the variable “perceived quality”. The Cronbach’s Alpha of the items was 0.947. Most scholars consider a Cronbach’s Alpha higher than 0.7 to be acceptable. Therefore, the scale mean of the perceived quality was computed and coded as “Total_Q”. In the following table (Table 2) the means and standard deviations of the variables are presented along with the correlations between them:
35 | P a g e Table 2: Means, Standard Deviations and Correlations
Variables Mean SD 1 2 3 4
1. Perceived quality 5.51 2.41 (0.947) 2. Willingness to pay 5.64 2.51 0.45**
3. Perceived price 6.58 3.07 0.4** 0.798**
4. Willingness to buy 5.5 2.67 0.305** 0.039 -0.07 -
** Correlation is significant at the 0.01 level (2-tailed)
5.1 Respondents’ characteristics
In this part of the paper the characteristics of the participants are presented in total and per group. A total of 240 respondents started the survey, but only 229 completed it, which means that 95% of the ones that started the online experiment completed it. The average age of the participants was 28 years, with the youngest participant being 19 years old and the oldest one 64. The demographics of the sample are further illustrated in the following table (Table 3):
Table 3: Respondents' Demographics (part 1)
Total (N=229) Condition 1 (N=55) Condition 2 (N=57) Condition 3 (N=58) Condition 4 (N=59) Gender Male Female 33.6% 66.4% 25.5% 74.5% 40.4% 59.6% 31% 69% 37.3% 62.7% Age 18-24 years 25-34 years 35-44 years 45-54 years 55-64 years 55% 20.1% 12.7% 10% 2.2% 52.7% 27.3% 9.1% 9.1% 1.8% 49.1% 24.6% 12.3% 14% - 62.1% 12.1% 12.1% 12.1% 1.7% 55.9% 16.9% 16.9% 5.1% 5.1% Income <750 RON 751 – 1500 RON 1501 – 2500 RON 2501 – 3500 RON > 3500 RON 12.7% 36.2% 21.8% 14.4% 14.8% 5.5% 29.1% 27.3% 20% 18.2% 15.8% 22.8% 29.8% 14% 17.5% 10.3% 50% 17.2% 12.1% 10.3% 18.6% 42.4% 13.6% 11.9% 13.6% Education Middle school Highschool Bachelor’s degree Postgraduate 1.3% 17.9% 60.3% 20.5% 1.8% 12.7% 61.8% 23.6% - 24.6% 54.4% 21.1% 1.7% 17.2% 65.5% 15.5% 1.7% 16.9% 59.3% 22%
36 | P a g e
Table 3 continued (part 2)
Occupation Student Employee Entrepreneur Unemployed Retired Homemaker 40.2% 45% 10.5% 1.3% 1.7% 1.3% 36.4% 50.9% 10.9% 1.8% - - 40.4% 42.1% 14% - - 3.5% 41.4% 46.6% 6.9% - 3.4% 1.7% 42.4% 40.7% 10.2% 3.4% 3.4% - Marital Status Married Single Widowed 26.6% 72.1% 1.3% 23.6% 76.4% - 29.8% 68.4% 1.8% 25.9% 70.7% 3.4% 27.1% 72.9% - Type of sugar Normal sugar Organic sugar 81.7% 18.3% 76.4% 23.6% 84.2% 15.8% 86.2% 13.8% 79.7% 20.3% The table above clearly illustrates that the females are greater represented in the total sample and in each group. The respondents were divided into five age categories:18-24 years, 25-34 years, 35-44 years, 45-54 years, and 55-64 years. Most of the respondents are in the first two categories in total and in each group, with the other age categories being under-represented. The means of the age in each group present no significant difference (Condition 1 – 28.18; Condition 2 – 28.88; Condition 3 – 28.79; Condition 4 – 29.10). The income was also divided in five categories and most of the respondents have an average monthly income between 751 RON and 1500 RON which is considered a low income in Romania, but the second category of income (1501-2500 RON) which is considered medium is also greatly represented in the sample. These results make sense considering the fact that 40% of the respondents are still students. Most of the respondents are employed though, and the other categories are under-represented. Furthermore, most of the participants of this experiment have a Bachelor’s degree or a Postgraduate diploma (80.5%) which means that the results cannot be generalized for the entire Romanian population, but only for higher educated people. Moreover, the majority of the sample is not married or widowed, but single.
37 | P a g e
5.2 Relationship between signals of quality and perceived quality
First of all it is important to mention that according to the manipulation check at the end of the questionnaire, not all the participants of the experiment observed the signals of quality. Only 77.2% of the respondents from the second group which were shown the package with the quality sticker (Appendix 6) noticed it and 87.9% of the participants in the third condition noticed the bio-label (Appendix 7). The respondents in the fourth condition are split into four categories: 44.06% of the respondents observed both signals (Appendix 8), 44.06% observed just the bio signal, 5.08% observed just the quality sticker and 6.8% didn’t notice any of the signals.
The first step in analyzing the effect of the independent variable “signals of quality” on the dependent variable “perceived quality” was to perform a One-Way Anova analysis used to compare the differences between groups. The results of the analysis are presented in Table 4. The Levene’s test was significant (p<0.05) which means that the variances across groups are not homogenous, but fortunately Welch and Brown-Forsythe tests results were both significant (p<0.05). Nonetheless, a method of post hoc test that does not rely on the assumption of homogenous variances was chosen, the Games-Howell procedure. Furthermore, the results show there is a significant difference between group means (p=0<0.05).
Table 4: Results One-Way Anova – Perceived quality
Dependent Variable: Perceived quality
Condition Sig. Games-Howell No signal Quality sticker 0.200 Bio signal 0.000 Both signals 0.000 Quality sticker No signal 0.200 Bio signal 0.000 Both signals 0.001 Bio signal No signal 0.000 Quality sticker 0.000 Both signals 0.967
38 | P a g e
The results indicate that there is no significant difference between the perceived quality of the respondents from the first group and the second one. That means that the quality sticker did not have a statistically significant effect on the perceived quality, F (3,225) =17.57, p=0.2 > 0.05. Furthermore, the Games-Howell post hoc revealed that there is a significant difference in perception of quality of the respondents from condition one compared to condition three and four (p=0<0.05), but there is no statistically significant difference between the perception of quality of the ones that were shown the package with the bio signal (condition 3) and the ones that were shown the package with both signals (condition 4). There is also a significant difference between condition two and condition four (p=0.001<0.05). The means plot reveals that the perceived quality is higher is in the second group compared to the first one and even higher in the third group. Therefore, the bio label is quite efficient in increasing the perceived quality, when another signal was added, the perception of quality actually decreased (Figure 5). The same analysis was performed once again only taking into consideration the respondents that noticed the signals according to the treatment of the experiment that they were exposed to and the results were similar. These results show that hypothesis 1a is partially supported because the presence of a signal of quality increased the perceived quality, but not all signals have this significant positive effect. Hypothesis 1b is also partially supported because increasing the quantity of signals did not have a significant effect on the perceived quality when compared to the bio label, but it did have a significant effect when compared to the quality sticker.
39 | P a g e
Figure 4: Means plot Perceived quality
The next step was to perform a Factorial Anova in order to test the main effect of each signal on the dependent variable perceived quality. The summarized results are shown in Table 5:
Table 5: Results Factorial Anova
Tests of Between-Subjects Effects Dependent Variable: Perceived quality
Source Sig. Partial Eta Squared
Corrected Model 0 0.19
Intercept 0 0.851
Quality Sticker 0.031 0.02
Bio Signal 0 .145
Both Signals 0 .129
The results of the Factorial Anova show that the model is valid (p=0<0.05) and that there was a significant main effect of signals of quality on the levels of perceived quality. The bio label had the highest effect on the perceived quality, p < 0.05, η² = 0.145. The quality sticker also had a significant effect on the perceived quality, but it was a low effect, p=0.031 < 0.05, η² = 0.02. When both signals were present, the effect on the perceived quality was moderate, p=0 < 0.05, η² = 0.129, higher than when just the quality sticker was present, but lower than when just the bio
40 | P a g e
label was present. Therefore, hypotheses 2a and 2b are supported because the relation between signals of quality and perceived quality depends on the type of signal and the bio label had a greater effect on the dependent variable than the other signal of quality (Quality sticker).
First of all, a One-Way Anova analysis was performed to compare the differences in willingness to pay between groups. The Levene’s test was significant again (p<0.05), but so were Welch and Brown-Forsythe tests (p=0<0.05). Therefore, the Games-Howell procedure was used this time as well. The results of the analysis are presented in Table 6:
Table 6: Results One-Way Anova - WTP
Dependent Variable: Willingness-to-pay
Condition Sig. Games-Howell No signal Quality sticker 0.823 Bio signal 0.000 Both signals 0.000 Quality sticker No signal 0.823 Bio signal 0.000 Both signals 0.000 Bio signal No signal 0.000 Quality sticker 0.000 Both signals 0.922
The results show there is a significant difference between group means (p=0<0.05). The post doc test revealed that there is no significant difference in willingness to pay between the first and second group (p=0.823>0.005), but there is a significant difference between the respondent’s willingness to pay in the first condition compared to the ones in the third and fourth condition. There is also a significant difference in the willingness to pay of the second group compared to the third and fourth group, but no significant difference was found between the third and fourth
41 | P a g e
group. The means plot (Figure 5) revealed that the willingness to pay is higher in the second (but not significant, p=0.823>0.05) and even higher in the third group (significant, p=0<0.05), but it decreased in the fourth group (not significant, p=0.922>0.05).
Figure 5: Means plot - WTP
Secondly, a hierarchical regression analysis was performed in order to study the relationship between the independent variable “perceived quality” and the dependent variable “willingness to pay”. This regression was used to explain/predict the effect of perceived quality on the willingness to pay after controlling for age, gender and sugar consumption per month. These control variables were included in the analysis to hold them constant for the calculations made about effect of the independent variable on the dependent variable. Therefore, this will help us come closer to examining the "true" relationship between the perceived quality and the willingness to pay which is independent of these demographic variables. The summary of the results of the regression are illustrated in Table 7:
42 | P a g e
Table 7: Results Hierarchical Regression – dep. Variable is WTP
R R2 R2 Change B SE Beta T Sig.
Step 1 0.209 0.044 0.044 0.018 Age 0.004 0.018 0.016 0.224 Gender -0.928 0.35 -0.174 -2.648 Sugar/mo -0.313 0.211 -0.105 -1.483 Step 2 0.52 0.271 0.227 0 Age -0.02 0.016 -0.079 -1.261 Gender -0.683 0.308 -0.128 -2.216 Sugar/mo -0.56 0.187 -0.187 -2.998 Perceived quality 0.521 0.062 0.501 8.351
In the first step of the hierarchical multiple regression, three predictors were entered: age, gender and sugar consumption per month. This model was statistically significant (F (3, 225) = 3.428; p = 0.018<0.05) and explained 4.4% of the variance in the willingness to pay. In step two perceived quality was entered as a predictor. This model was also statistically significant F (4, 224) = 20.793; p < 0.01 and the total variance explained by it as a whole was 27.1%. The introduction of perceived quality explained an additional 22.7% variance in willingness to pay, after controlling for age, gender and sugar consumption per month (R2 Change = .227; F (1, 224) = 69.747; p = 0<0.001). In the final model three out of four predictor variables were statically significant, with perceived quality recording a higher Beta value (β = 0.501; p<0.01) than gender (β = -0.128; p = 0.029< 0.05) and sugar consumption per month (β = -0.187, p = 0.003 < .01). Therefore, hypothesis 3 is supported because the results of the hierarchical regression showed that the perceived quality has a positive effect on willingness to pay.
The moderating effect of income on the relationship between the perceived quality and willingness to pay was also tested. In order to test this effect, the computational tool PROCESS was used (Model 1). The summarized results of the analysis are presented in Table 8: