Turning credence attributes into search attributes.
How consumers react on the provision of additional information on product labels.
***
- ABSTRACT -
A research was conducted to find out how the provision of information that supports credence claims made by product labels affects the perception and behavior of those labels and products. 920 panel-members of an on-line auction platform placed a bid and answered several questions after being shown one of five different treatments of an
advertisement for a 5 liter can of organic olive oil. The treatments differed in amount and nature of information and as such were hypothesized to show differences in respondents' willingness-to-pay and perception of organic nature of the product. The analyses did not show any significant results in terms of differences between the treatments. As such there was no proof found that information supporting credence claims of a product label has a positive effect on the
perception and willingness-to-pay of the consumer. Proof was found that the level of trust towards the label and the perception of organic nature of the product positively influence consumers’ willingness to pay.
***
Author: Splinter Dreesmann, 6134041 Supervisor: Nachoem M. Wijnberg
A word of thanks
As the final chapter of my somewhat lengthy career as a Business Studies student at the University of Amsterdam, I started working on this thesis 6 months ago. But as they say, time flies when you’re having fun, and as such today means the end of this beautiful last period. While writing this thesis I was studying at Bocconi University in Milano, Italy. I must acknowledge that the beauty of the city, the people and the food have certainly inspired, and distracted me throughout the process. There are lots of people that I owe thanks to. To start with Nachoem M. Wijnberg, for his
supervision and the possibility to use olive oil as the product for auction. Bram Kuijken and Anouar El Haji for their help and for setting up their great platform Veylinx. Joris Ebbers, for managing the great EMCI track and for offering me the possibility to spend the last semester of my study at Bocconi in Milano. Jordi van den Berg for being a great and inspiring friend. Emanuele, Phil, Andreas and Henrik for teaching me about life and especially about myself. Lombardo, Loop, Rookies, Louis van Gaal, and of course last but not least, my family for their never-ending support.
I hope you enjoy reading my thesis.
Table of Content
1. Introduction ... 4
2. Literature Review ... 9
2.1 Information Theory ... 9
2.2 Product Labeling... 11
2.3 Consumer attitude and willingness to pay for labeled products ... 14
2.4 From credence to search attributes ... 15
2.5 Organic ... 17
2.6 Organic as a credence attribute... 18
2.7 Organic labels ... 19 2.8 Authenticity ... 23 3. Methodology ... 25 3.1 Research Design ... 25 3.2 Data Collection ... 26 3.3 Treatments ... 28 4. Results ... 31 4.1 Descriptive statistics. ... 31 4.1.1 Sample characteristics. ... 31
4.1.2 Descriptive statistics of the treatments and willingness to pay. ... 33
4.2 Correlations ... 39
4.3 Hypotheses ... 41
4.4 Alternative findings ... 46
5. Discussion ... 48
5.1 Discussion of the Results... 48
5.2 Limitations ... 52
5.3 Future research ... 53
6. Conclusion ... 54
7. References ... 56
1. Introduction
In recent years, there has been a growing attention in scientific research as well as in popular media directed towards questions surrounding health, safety, social and environmental effects in the food industry (Hughner, McDonagh, Prothero, Schultz II and Stanton, 2007; Tagbata & Sirieix, 2008). In particular, a lot of attention is directed towards the agricultural industries, which have been a
continuous subject of discussion due to their ongoing process of technological change over the past 150 years (Hobbs & Kerr, 2006). This technological change has cut the necessity for labor
dramatically and has also facilitated the development of the mass farming facilities needed to feed the ever growing and ever more consuming population of the world. The implications of these developments are ambiguous to say the least. For one, the exit of traditional farmers in rapidly developing countries has shown to be a problem for governments and policy makers in terms of international trade, subsidies and protectionism, with its obvious implications for social systems (Hobbs & Kerr, 2006). But these developments have also put in place a growth driven food supply system that has in recent years sparked more and more questions and discussion about health issues, food safety, animal well-being, production processes, the use of genetically modified organisms, pesticides and many, many other issues (Grunert, 2005).
In general, issues concerning the effects that products have on consumers and the environment are found throughout many types of different industries. In addition to food industries, such issues are found for example in pharmaceutical industries, forest processing industries and energy providing industries (Teisl, Peavey and O'Brien, 2001; Bird, 2003). The binding factor of these types of issues throughout all different industries resides in the fact that in general the consumer is not able to evaluate these long-term effects of the alternative products offered to him in the market place. And as such, the market will not function in the sense that it will “punish” those producers that produce malevolent products. This problem is in part explained by the economics of information, a concept developed by George Stigler (1961), that describes the difficulties one stumbles upon when
searching for information to evaluate the price and quality of goods. The work done by Stigler (1961) was later used for the development of Information Theory (Nelson, 1970, 1974; Darby & Karni, 1973). Information theory states that the information concerning a certain product can basically be categorized into three types of attributes: Search, experience and credence attributes. Search attributes are those characteristics of the product that are used by consumers to evaluate the product before consumption. For example, when a consumer decides to buy a new t-shirt, the color, the design and the price of the shirt are important search attributes which will for a large part
determine which shirt the consumer will buy. Experience attributes are those product characteristics that can only be used by consumers for evaluation after the product has been consumed. In the example of the t-shirt, the consumer may evaluate his purchase after a while considering the fit of the shirt after he has used it for a while. Whenever the consumer is unhappy with certain aspects of his purchase, in this case it might be that the shirt loses its fit after washing it a few times, the consumer will take this information with him when making his next purchase decision. Credence attributes are characteristics of the product that are very difficult for consumers to evaluate before and even after the consumption of the product. Here one might think of the conditions under which the t-shirt was produced, and the effects on all manufacturers and suppliers in the production chain (Nelson, 1970; Darby & Karni, 1973). Now when regarding search and experience attributes, it is very likely that the market will eventually filter out malevolent producers, because products that don't appeal to consumers before, and/or after purchase are not likely to be bought again, however low the price of the product may be. As such, there is little or no incentive for these producers not to try to fulfill the needs of the consumers as well as possible. However, since credence attributes are very difficult, costly and timely for consumers to evaluate, the demand side of the market is not able to filter out products that are no good. A common, and more and more widely applied attempt to counter these types of information problems is found in the use of product labels. A very common example which lies at the foundation of this research, is the organic label. The organic label claims
that a product was produced in an organic way, and as such, the label helps the consumer to
distinguish it from other products that were not produced in this way. This is basically the function of all types of product labels found throughout all types of industries. To give the consumer a means to distinguish products which are for example environmentally friendly from those that are not. However, these types of beneficial attributes that are claimed by such a product label, very often mean that a product is also more expensive for consumers to buy. When looking at the organic food industry for example, the price premium that has to be paid for organic products is the number one reason for consumers not to engage in buying organic food (Hughner et al. 2007). This problem is more complicated than it seems though. Because for a large part of consumers the problem is not that they do not want to pay a bit more for products that have beneficial effects on themselves or on the environment. The problems lies more in the lack of knowledge and a general sense of distrust concerning the institutions that are responsible for the product labels (Caswell & Madjuzska, 1996; Hughner et al. 2007). This distrust and lack of knowledge is understandable, if we take in
consideration for example the fact that in recent years a lot of research has been conducted towards finding out if and how organic products are actually more beneficial for the consumers health then the conventional products (Lairon, 2009; Lima & Vianello, 2010). Moreover, the consumer
confusion might even be reinforced when large multinational companies now begin to capitalize on the organic food trend by introducing their own organic product lines in addition to their
conventional product lines (Hughner et al. 2007).
Regarding this problem, research has come to the insight that an institution which is responsible for product labels should best be a public institution (Caswell & Madjuzska, 1996). Public institutions are regarded as independent and generally more trustworthy than their privately owned
counterparts. Another advantage of public institutions is that they have access to regulatory power, meaning that they can actually set guidelines and rules and punish those that do not abide them.
Research suggests that the development of more concise, more standardized and more transparent labeling strategies may lead to a stronger relationship between labels and their potential consumers (Loureiro & McCluskey, 2000). A concept which may aid in the facilitation of such a relationship is the provision of information surrounding the nature and meaning of the label. In other words, explaining to potential consumers what the label exactly means and why consumers are asked to pay a higher price for products bearing the label. It is important here to note that this provision of information may in large part already be available, however it is often very costly or time
consuming for consumers to search for this information. To make the information more accessible to the consumer, providing an organic label with clear and transparent information about the benefits of an organic product and the reasons why this certain product has been labeled as such, may help turn these credence attributes into more search like attributes, which can be evaluated by the consumer before consumption. The provision of such information may help to take away some of the vagueness and skepticism surrounding the label and may also turn out to be an effective way to educate the consumer. Regarding the provision of information on product labels, Grunert & Wills (2007) concluded after an extensive review of existing literature, that no clear understanding is present as to what actually goes on in the consumers mind when regarding such information. This research aims to find out whether this provision of additional, on the spot information, will help change the perception of consumers about products with a product label, and in turn if the label that is supported with this extra information will actually lead to a higher consumer willingness to pay than products that lack this additional information. The research question around which this research is designed is as follows:
What is the effect of information concerning product labels with credence claims on the perception and behavior of consumers regarding products which are for an important part defined by these credence qualities?
behavior concerning organic labels by the provision of information, a quantitative research will be conducted using data gathered with an on-line auction model. This on-line auctioning system allows for the measurement of consumer willingness to pay, which will serve as a means to measure
consumer behavior, and also it allows for the measurement of consumer perception through the use of a series of closed questions. Veylinx, as the auction system is named, allows for multiple
treatments of products, which enables us to research what the consumer differences are between treatments that differ on the point of degrees of information provision. A more detailed description of the auction system and the further methodology of this research can be found in paragraph 3. The subject of this particular research, and thus the means through which consumer behavior and perception will be measured, will be organically produced olive oil. The advantage of using olive oil for this type of research is that it is a product that has become widely known and used in the past decades. Olive oil is a product that is widely available in the country of research, the Netherlands, and it is available in a large range of different prices and qualities. This makes it a product for which consumers have a large amount of freedom in their choice, and as such it is well suited to investigate in precision how consumers perceive organic labels and what the influence of information provision is on these consumers.
In the next paragraph, paragraph 2, more light will be shed upon the literature on which this research is based, including literature on information theory, literature on product labels and literature on organic products. In paragraph 3, the methodology, the research design and the operationalization of the research will be presented. In paragraph 4 the results of the research will be presented, which will then be discussed in paragraph 5, together with the limitations and opportunities for future research. In paragraph 6 the conclusions of this research will be presented.
2. Literature Review
The aim of this study is to investigate in what way labels on food products influence the behavior and perception of consumers regarding certain products. In particular, this study is meant to reveal how and if the provision of additional information supporting labels that reflect a certain beneficial claim, which can’t be evaluated by the consumer on site, changes the perception of the consumer, and in doing so will affect consumer behavior. In this particular study, a research will be conducted towards organic labels. In general, research towards consumer behavior and perceptions
surrounding certain product characteristics is often based on the framework of information theory. As such, this section will first serve to investigate further the properties of information theory and its relevance for the case of product labeling.
2.1 Information Theory
George Stigler (1961) was among the first to coin the importance and impact of information to daily economic transactions: “The identification of sellers and the discovery of their prices are only one
sample of the vast role of the search for information in economic life” (Stigler, 1961). This basic
idea was in turn used by Phillip Nelson (1970) to develop a framework on which consumers base their evaluation and perception of the quality of products. Nelson (1970) uses Stigler’s broad idea on information economics in a somewhat narrower perspective. Nelson’s work departs from the assumption that consumers already know where the alternatives that are to be taken into
consideration are available to them. Then, the information problem faced by these consumers, is restricted to evaluating the price and quality of each of these alternatives in order to make their decision. In order for these consumers to make an adequate decision, Nelson identifies two types, or two parts of information processes, that are used by consumers to evaluate their options. The first
part of the information process is identified by Nelson as the search process, which is restricted and defined by two simple propositions: a product or an alternative option belongs to the search process if (1) this option is inspected by the consumer, and (2) that inspection takes place prior to the purchase of the product. The second part of the information process is identified as the experience process. This is the part that is defined by the consumer’s evaluation of the product after purchase (Nelson, 1970). Nelson (1970) basically identifies a trade-off between these two parts of the
information process based on price. For an expensive product, the information process will exist for the largest part out of search, because an adequate and intensive evaluation of the available
alternatives prior to purchase will be preferred over the actually buying and trying several options since this will imply a waste of money. For very cheap products however, the costs inclined by extensively searching for the best option may not weigh up to the extra costs inclined by simply buying a few alternatives and testing them through experience (Nelson, 1970). This distinction between search and experience characteristics of products was complemented by the work of Darby & Karni (1973). Darby & Karni (1973) introduced into the framework the term “credence”
qualities, defining them as those qualities claimed by producers that are even after purchase very hard and very costly to evaluate. An important notion that is mentioned by Darby & Karni is that due to the existence of a general lack of knowledge residing under consumers, and the generally relatively high amount of additional costs that consumers must incur in order to fully evaluate these credence qualities, there is an opportunity and a temptation for producers to enter into fraudulent activity. This remark, which is inextricably linked to the nature of these credence qualities, is also an important remark which lies at the foundation of this research; consumers’ lack of knowledge and capability which is needed to check the truthfulness and effects of the beneficial claims made by producers of products which bare an organic label can result in skepticism and distrust towards these producers.
information, since the consumer does not have the information that is possessed by the producer before he or she buys the product in question. In the case of experience attributes, the problem of imperfect information will for a large part be solved by reputation. Since the consumer can assess the quality of a product after purchase, and will base his or her future purchase decisions for a large part on this assessment of the experience attributes of the product, there is little or no incentive for producers to use the information gap in a way that is harmful for the consumer. In this specific case of experience attributes in the food industry, the imperfect information problem will very likely be solved within the market itself, since the main experience characteristic that will be used by consumers to evaluate producers is the perceived quality of the product. However, in the case of credence attributes of food products, the imperfect information problem is much more complex. In the context of food, credence attributes could be anything ranging from benefits for the consumer's health, to environmental benefits. The main factor that makes these attributes credence attributes is that it is practically impossible for the consumer to evaluate the quality of the product in terms of these supposed benefits. Even by actually consuming the product, these attributes will not unveil their full effects to the individual consumer and as such the market cannot control for cases where information provided by producers may be misleading.
2.2 Product Labeling
In this particular case of imperfect information problems due to products with credence attributes, research has concluded that there is an important role to be played by governments, in the function of trustworthy and reputable certification agents. The certification labels which are required by policy of governments for products that make a certain “credence” claim can help consumers in making their evaluation of the quality of those products (Caswell & Madjuzska, 1996). It is
necessary that these certifications are set up by public institutions, and not private institutions, since public institutions have regulatory power and a higher degree of credibility for the consumer.
When regarding the literature on product labels, one notices that the practice of labeling touches upon a wide number of industries and applications. For example, Bird (2003) explores the status quo of the energy labeling practices in the United States. He concludes that the certificates as required by different states seem to enable consumers to make reliable comparisons between suppliers and help consumers in educating themselves as to what the actual effects of choosing for green energy are. In this case, the certificates or product labels seem to help overcome the problems that arise due to the imperfect information situation.
In a research towards the certification and labeling of forest products, Teisl, Peavey and O'Brien (2001) note that labels that just mention the “environmental goodness” of a product seem to be less effective as labels that actually provide the consumer with insights about the product's
environmental characteristics. Here, the authors make a clear distinction between just mentioning the credence attribute of the product, and actually supporting this credence claim with
characteristics of the product. More focused on the food industries, where labeling schemes affect the environment as well as the consumer's personal wellbeing, there are two main important directions of research. The first direction of research is concerned with front-of-pack nutrition labeling. When regarding this type of labeling, Grunert & Wills (2007) have conducted an extensive review of the literature. In their synthesis of earlier research, Grunert & Wills conclude with four important conclusions on the way consumers perceive and act towards this front-of-pack labeling. A first important point emphasizes that consumers generally understand the relationship between food and health, and as such there is a widespread interest for nutrition information on food packages. A second conclusion states that front-of-pack nutrition labeling is most valued by the consumer in a simplified form. A good example of simplified information is the so called traffic light nutrition label, which uses color coding to show consumers the levels of certain nutritional attributes in a certain product. The third conclusion presented by Grunert & Wills states that most consumers indicate to believe to be able to understand and reproduce key information presented to them
through use of a label. This has only been tested in an experimental setting however. This relates to the fourth conclusion, which states that there is actually very little insight into how consumers use labeling information in a real world situation. Grunert & Wills also compare the results of their study to an earlier review conducted by Cowburn & Stockley (2005) who, like Grunert & Wills (2007) emphasize that the importance of understandable labeling is key in attracting consumers to actually look at and process the information on the label. As such, labels should not be too complex, should not contain to many technical terms, and should refrain as much as possible from the use of numerical calculations and for some consumers even percentages (Cowburn & Stockley, 2005; Grunert & Wills, 2007).
The second important direction of research regarding the labeling of food products is the research concerning organic labels. Like other types of product labeling, organic certificates deal with issues of understanding and trust of the label (Hughner et al., 2007). In a research towards consumer perception and attitudes towards different formats of organic labelling schemes, Janssen & Hamm (2012) have found that consumers prefer and have more positive perceptions and attitudes towards those labelling schemes that have a higher rating in terms of awareness, trust, credibility and standards and control systems. However obvious this may seem, it does stress clearly the importance of a centralized and well-known certification agency with clear standards and
transparent control procedures. However, Janssen & Hamm (2012) do indicate that this third party certification does not automatically overcome the problems that arise due to the credence nature of the product attributes relevant for the organic nature of the product. According to Janssen & Hamm (2012), in order for an organic certificate to be successful, not only needs the consumer awareness of the logo be high (so the recognition of a logo as part of an institutional organic certificate), also the consumer needs to have knowledge and consequent positive attitudes towards the program that underlies the certificate.
subject seem to agree on several points. First, it seems that certificates and other labeling programs can help consumers in evaluating and differentiating between several available products. However, it seems of great importance for a product label to effectively work, that the information provided is easy to find and easy to process. Furthermore the institution that hands out the label should be known and trusted, which can be reached by using centralized and transparent labeling programs.
2.3 Consumer attitude and willingness to pay for labeled products
The above mentioned examples of product labeling programs throughout different industries have given an insight into what is necessary for product labels to actually work. These evaluations are based on the prerequisites that determine when the consumer has the most positive perceptions and attitudes towards these labels. Coherent with the nature of credence attributes, it remains the question however, in what way consumers with positive perceptions and attitudes towards certain product labels will actually behave in the market place. When they actually have to pay a higher price for the labeled product when compared to a non-labeled alternative. As is emphasized by Hughner et al. (2007) in the specific case of organic labeling, the price premium is one of the major deterrents for consumers to actually buy organic labeled products. This specific problem makes it interesting to investigate the relationship between the perceptions and attitudes that consumers have and their actual behavior. A commonly used way to study consumer behavior in the case of product labels is the measurement of the consumer's willingness to pay for different products.
In order to understand this relationship between the product label and the consumer’s willingness to pay for the product bearing the label, it is important to understand that, according to the information theory, these labels serve as some kind of quality signal for consumers. As such, understanding the economic effects of product labels in the markets they operate in, is actually explainable by the information theory developed by Nelson (1970) and Darby & Karni (1973), as presented earlier. The product label serves as a signal for the consumer about the credence qualities of the product. As
these credence attributes are not, or barely, measurable for the consumer, the consumer's
willingness to pay for the product with a label is dependent on the degree in which the consumer understands and trusts the label and its owner (Caswell & Madjuzska, 1996). It seems arguable that a way of getting consumers to pay a price premium for labeled products with credence attributes is to try and turn these credence attributes into search attributes.
2.4 From credence to search attributes
One thing that is clear is that the key to understanding and control of the producers and certification agencies involved in handing out these labels of credence qualities is transparency. This
transparency can be introduced in a consumer friendly way by offering the consumer the possibility to evaluate on the spot whether or not a product fits the consumer's ideas as to what an organic product should or should not be. This would have as a result that transparent and standardized labels are most effective in building positive search attributes for consumers. Obviously the status quo in the industries concerned at the moment is a such that there are several labels already available which offer a standardized control on what the products that bear those labels have as organic characteristics. However, a lack of trust, and a lack of knowledge and a high price premium
combined with a general distrust of large organization often leads to purchasing decisions that keep the status quo of mass consumption companies in mass consumption industries alive. Moreover, even though in the case of several labels some information about credence characteristics is already disclosed, many consumers will have difficulties in evaluating this information nonetheless. This is specifically the case when there resides a discrepancy between the consumers availability of time and access to specific knowledge and skills on the one side, and the time, skills and knowledge needed to accurately evaluate these credence characteristics on the other side (Wynne, Benton & Redclift, 1994). A well-designed label could serve as an important cognitive support that deals with these consumer constraints in an effective way (Valceschini, 1999). Several authors note that such a well-designed label would typically succeed in transforming credence attributes into search
attributes (Caswell & Madjuzska, 1996; Loureiro & McCluskey, 2000). This would in turn aid consumers in making informed choices concerning these credence attributes based on the provision of easy to process, reliable information (Loureiro & McCluskey, 2000). As such, this system of information provision can be used as a regulatory technique which gives consumers the possibility to access a sufficient amount of information in order to protect themselves against unsafe products and unfair seller behavior (Caswell & Madjuzska, 1996). However, as was mentioned by Grunert & Wills (2007) there is very little insight in how consumers actually use the information provided on food labels. Here lies an important literature gap which this research tries to close. By measuring behavior and perception of a product in an experimental market place, this research aims to find out how consumers actually use information on product labels, in terms of evaluation of the label and buying behavior.
Summarizing, there are several factors identified in prior research that can help in building a successful labeling scheme. First of all, the institution that gives out the certificate or label should be trusted by the consumer. Only in this way will the label have an actual advantage over labels that are set up by private institutions. Second of all, labels that provide consumers with information supportive of the (credence) claims of the label are more positively assessed then those labels that just make credence claims. However, it is important that this information is structured in such a way that it respects the knowledge and time scarcity of the consumer. This basically means that the information on the label should not be difficult to understand and should be provided in a way that is quick to process. The final point made in the literature is that the certificates should be built up out of clear and transparent standards and control measures, making it easy to understand what is actually demanded of producers and what is actually done by the certification agency to control the producers.
2.5 Organic
When exploring the properties of the empirical setting of this research, it is valuable to start with formulating what the term organic actually entails. Now, although the term organic is certain to evoke some sense of healthiness and environmentally friendly production with consumers, authors only rarely record a definition of the term organic in their articles. Whether or not a product is to be labeled organic, obviously depends on whether it was produced in an organic production system. Lampkin (1988) was among the first authors to include the definition of organic production as formulated by the United States Department of Agriculture in his work: “A production system
which avoids or largely excludes the use of synthetic compounded fertilizers, pesticides, growth regulators, and livestock feed additives. To the maximum extent feasible, organic farming systems rely upon crop rotations, crop residues, animal manures, legumes, green manures, off-farm organic wastes, and aspects of biological pest control to maintain soil productivity and tilth, to supply plant nutrients, and to control insects, weeds, and other pests” (USDA, 1980 as found in Lampkin,
1988). This definition would later on become commonly used and it has been widely adopted in prior research (Browne, Harris, Hofny-Collins, Pasiecznik and Wallace, 2000). In their research towards the marketing and pricing strategies of organic producers compared to producers of
conventional products, Geier & Vogtmann (1984) have noted that it is possible for producers to set prices at a premium level for several market and product segments. In line with this, Batte, Hooker, Haab & Beaverson (2007) have also found that magnitudes of willingness-to-pay-premiums vary significantly among consumers groups making targeted marketing an effective strategy for marketers of organic products. Various researchers have noted however, that although most consumers will have a certain feeling with, and an idea about the term “organic”, most consumers are not familiar with the standards and practices surrounding organic food production (Hughner et al., 2007). This lack of knowledge may also have its negative effects on the behavior of consumers regarding food that is labeled as organic, as is stated by Hughner et al. (2007) in their extensive
review of prior research on the motives of consumers regarding organic food. As it turns out, skepticism of certification boards and organic labels is one of the most important deterrents of consumers to engage in buying organic food, together with high price premiums and a lack of availability (Hughner et al. 2007).
Zanoli, Scarpa, Napolitano, Piasentier, Naspetti and Bruschi (2013) have studied the influence of different types of cues on the preferences and willingness to pay of Italian consumers for organic, conventional and genetically modified products. Zanoli et al. (2013) found that credence cues appeared to be more important than search cues in explaining the consumers’ willingness to pay for organic products. This is explained by the notion that consumers will probably evaluate products roughly the same in terms of their search cues. However, consumers may differ highly in their values and beliefs, which can in turn lead to large differences in preferences which can be satisfied with credence attributes.
2.6 Organic as a credence attribute
Earlier research has already identified the characteristic of organic to usually be a credence attribute since it is very difficult for the individual consumer to evaluate the organic nature of the product (Poelman, van Loon & Dicke, 2008; Tagbata & Sirieix, 2008). The main core of an organic label’s claim resides in the way the product was produced. More and more consumers derive utility from buying and using these products which were produced under specific processes, for example because they entail environmental benefits or benefits for the consumer’s health (Grolleau & Caswell, 2006). However, in the nature of these credence attributes lies that they are very difficult to check and evaluate by the consumer himself. Think for example of the cost inclined for a single consumer to check whether or not his consumption of organic products has had any positive effects on his health. Or think about the lengths that a single consumer must go through in order to find out what the exact environmental impact is of a company that produces organic food, and how this
compares to all companies that offer the same type of nutrition, with or without an organic label. This difficulty is illustrated further by the fact that some institutions, like the EU organic certificate, not only include the production process of the producer in their evaluation, as is suggested by the definition of organic by Lampkin (1988). For example, the EU certificate also includes an
evaluation of the companies supply chain, its transport system and its Co2 emissions. Consequently, this might lead to a situation in which the average consumer, who may or may not have a positive evaluation of the credence attributes as presented by the organic label, will be negatively influenced by the fact that he or she cannot determine for certain that the product’s credence attributes will actually produce benefits to the extent that the price premium demanded for the product will be worthwhile.
2.7 Organic labels
As was stated earlier, prior research has identified that product labels handed out by public
institutions are a possible method to overcome the problems inclined by the presence of information asymmetry when regarding product labels that make a credence claim. Since the empirical part of this research is focused on organic labels, the specific organic label which will be used for this research is the E.U. Organic Logo. Information as to what this label which is issued by the European Commission actually means for the producers and their products is easily accessible on the website of the European Commission (ec.europa.eu). However, one will only stumble upon this information when one actively searches for it, and for daily activities like grocery shopping,
consumers do not have the habit of engaging in an information search before going off to their local grocery store. Off course, the rise of the Internet has influenced the accessibility of information and consumers are much more powerful in their evaluation of product characteristics than they were some years ago, but technology has yet to bring about efficient changes in the direct environment of decision making which will help consumers make more informed decisions. So, although many
consumers may not know about the information that is readily available on the website of the EC, it makes very clear what the exact standards and guidelines of the label are, and thus it will be very valuable to get a more detailed insight in what this label actually means. Not only in order to gain more insight into what the concept of organic actually means and how this is executed by the EU, but also in order to determine which useful parts of this definition of organic can be used to turn the credence characteristics of the organic label used in this research into search characteristics.
A glance towards the contents of this EU organic logo (see appendix Y) leads to the quick conclusion that there is a very complete policy attached to it. Not only in terms of subjects the organic logo touches upon (health, animal welfare, sustainability, local production, exemption of GMO’s and strict limits on the use of chemicals, antibiotics and additives etc.), but also in terms of risk assessment and control programs. It seems hard to believe that a consumer who recognizes the EU organic logo, and knows all these guarantees and policies attached to it, would not be willing to pay a price premium for the product in question. However, the problem here may still be that the consumer is not in a financial position to pay the demanded price premium or doesn’t trust the certification agent (in this case the EU).
But, as Hughner et al. (2007) noted, the biggest gap between beliefs and behavior surrounding organic labels resides in the general consumer’s lack of knowledge about the exact standards and practices that surround organic food production. In other words, most consumers do have a certain feeling with the credence attributes of organic products, but they lack the knowledge to exactly specify what those credence attributes actually mean, and thus they might fail to give sufficient weight to these attributes at the moment of purchase decision. Providing the consumer with
information that turns these credence attributes into search attributes, which can help the consumer to actually understand and evaluate these attributes at the moment of the purchase decision, may thus have a positive influence on the consumers perception of the organic product, and may also positively influence the consumers behavior.
Concluding, it seems that the credence characteristics of the (in this case) EU organic logo can be broken down into exact information, and that these pieces of exact information can in turn be used as benchmarks through which the consumer can learn about the standards, practices and benefits that the organic product under his or her consideration has.
Hypotheses:
As was mentioned in the literature review by Grunert & Wills (2007), research has yet to unveil what actually goes on in the consumers mind when dealing with the information on product labels. There is no clear image of the process that goes on in the “black box” of the consumer, and as such it remains difficult to understand what kind of information, what volume of information and what form of information are effective for different kinds of purposes. As an attempt to find out how consumers deal with information on product labels in market situations, the hypotheses in this research were designed to explore in what way information provision in the form of simple text influences perception and behavior of consumers towards products with credence attributes, in our case organic olive oil.
The idea that information can have an effect on consumers perception and behavior in this specific case, is based on the notion by Hughner et al. (2007), that the largest part of consumers are not familiar with the standards and practices that surround organic food production. Consumers lack the knowledge and information needed to evaluate the credence attributes of organic products, and as such the provision of such information could help consumers in evaluating these credence attributes and consequently making more informed choices (Loureiro & McCluskey, 2000). Furthermore, as was found by Teisl, Peavey and O’Brien (2001), a product label seems to be more effective when it not just mentions the credence attributes, but also provides information about relevant
characteristics. Thus, the provision of information is hypothesized to influence the behavior of consumers, leading to the first hypothesis:
H1: When a product label is fitted with information that helps consumers to evaluate the credence claims made by that label, consumer willingness to pay for that product will be higher.
At the foundation of the consumer’s perception and behavior towards a product with credence attributes lies the consumer’s evaluation of these attributes. Although we don’t know much about what happens in the consumers mind when regarding a product label and how this influences his or her perception and behavior (Grunert & Wills, 2007), in this research it is hypothesized that when more information is presented to the consumer, the positive influence on perception and behavior will be larger. Furthermore, it is hypothesized that the perception of the relevant credence attributes is a mediator for the relation between the information on the label and the consumers behavior (willingness-to-pay). These assumptions are supported by the idea that providing more units of information will make the credence attributes more measurable for the consumer, and as such the degree of understanding and trust towards the label and the producer will be positively influenced, which in turn should have a positive influence on consumers’ willingness-to-pay (Caswell &
Madjuzska, 1996; Loureiro & McCluskey, 2000). Building on this, hypothesis 2 & 3 are formulated as follows:
H2: When more units of information will be shown accompanying a product label, this will positively influence consumers’ perception in terms of the credence claims of the product.
H3: A respondent's overall more positive assessment of the credence attributes of a product bearing a product label will lead to a higher willingness-to-pay.
2.8 Authenticity
When regarding the above described way in how products with a credence label can overcome the problems of imperfect information for the consumer, it is important to evaluate other options as well. In the particular case of overcoming issues of trust, awareness, credibility and standards surrounding the claims of beneficial production for the consumer that a product and its label make, instead of emphasizing the characteristics of the institution that is responsible for the label, the producer itself could also try to make certain credence attributes more into search attributes. In the case of organic products, there is a rather high and relevant comparison that can be made with the authenticity of a product. The fact that there is not one clear definition of what authenticity actually entails, makes it an especially interesting subject to study when regarding its effects on consumers. Authenticity can be perceived in different ways in different contexts (Boyle, 2003), and there may also exist discrepancies on what authenticity means for consumers and producers within the same industry (Holt, 2002). It may well be that the different perceptions of defining authenticity may in part form a foundation for the price premium that consumers are willing to pay. In order to
investigate this, it is very helpful to come up with a definition of authenticity that clearly
demarcates what it is expected to mean in particular industries. Some authors answer the problem of defining the concept of authenticity with relatively simple solutions. For example, Lunardo & Guerinet (2007) define authenticity as “being original, or being faithful to an original”, which is rather similar to a definition provided by Beverland & Farrely (2010), who define authenticity as “what is genuine, real, and/or true”. Although in general the reader will have little trouble in understanding what these authors mean with terms like original and genuine, these simplified definitions seem to just shift the problem of defining authenticity, rather than giving an applicable and comprehensive definition. In our case, a more useful definition is provided by Michael
Beverland in his earlier work studying the attributes of authenticity for luxury brands of wine, where he defined authenticity as “a story that balances industrial (production, distribution and
marketing) and rhetorical attributes to project sincerity through the avowal of commitments to traditions (including production methods, product styling, firm values, and/or location), passion for craft and production excellence, and the public disavowal of the role of modern industrial attributes and commercial motivations” (Beverland, 2005). Outside of this definition, some other authors
have come up with some interesting ideas regarding the concept of authenticity that may be rather useful in investigating the relationship between authenticity and organic nature. For example, Peterson (2005) mentions that authenticity is “a claim that is made by or for someone, thing, or performance and either accepted or rejected by relevant others” (Peterson, 2005, p. 1086). This would imply that when a producer claims that a product is authentic, it will only have beneficial effects for the producer when this claim of authenticity is accepted by consumers. This very nicely fits with what Grayson & Martinec put forward: “what is perceived as authentic must conform to consumers’ mental frames of how things ought to be” (Grayson & Martinec, 2004, as found in Beverland, 2005b, p. 251). These notions of authenticity have clear relations to the credence character of organic products. In both cases there are certain claims made by producers that are in turn to be evaluated by the consumer, who typically has a shortage of information to properly evaluate the information. Beverland (2005) notes that the perception of authenticity is typically build upon objective and subjective elements. Objective elements may for example include intrinsic product elements, the production process and links to place and historic style, whereas subjective elements are created by firm members, consumers and other stakeholders and may include stories or myths about the product and claims about status (Beverland, 2005). This distinction shows great similarity to the distinction between search and experience attributes on the one hand, and credence attributes on the other hand. In order to overcome the problem of conveying in a credible way the subjective elements of the authentic claim, Beverland states that there is an important role for the emphasis on the objective elements. In other words, measurable information should be used to help make the unmeasurable more credible. This is coherent with what this research aims to investigate
when regarding product labels. This leads us to the reason the term of authenticity was introduced in this paragraph. Instead of subjecting itself to the standards and controls of a centralized and public certification institution, a producer might decide to counter the problems of credibility, trust and production standards by identifying himself as an authentic producer. Consumers are more likely to trust the producer and his claims when they can attach a story, a face and a location to the product (Beverland, 2005). Moreover, this type of information may even prove to be more valued and easier to process by consumers than more technical and numerical information on product characteristics. The described relationship between the claims that come with the organic nature of products and the authenticity characteristics of products leads to the fourth hypothesis:
H4: Mentioning information on the authenticity of the product and the production methods will serve as an alternative to make credence attributes into search attributes, through giving the producer an identity, and will thus lead to a higher perception of the organic nature of the product.
In order to find an answer to the hypotheses that are formulated and consequently to find an answer to the research question that lies at the foundation of this research, in the next paragraph the specific empirical setting of the research will be explained.
3. Methodology
In this section, the empirical setting of the research will be described. This will be done through an explanation of the research design, the data collection, the sample and the variables.
3.1 Research Design
liters of olive oil. The aim of this research is to investigate whether, and if so in what way, the presence of information influences the perception and behavior of the consumer towards a product with an organic label. The choice for olive oil as the object of auction is justified by the fact that olive oil is an ordinary product, suitable for daily use. As such, there are less implications in terms of generalizability of the eventual findings then when a product was used that was less accessible for the population and the sample. Another important aspect is of course that olive oil can be either organic, or not. The product itself suits the auction method well. Using several advertisements that are treated in different ways, the decision making process that normally takes place in a
supermarket or other shop can be relatively well simulated. The advertisements used in the auction, are roughly the size of the front of the respective olive oil cans.
Experiment
Since the objective is to research the influence of different types and different amounts of information on the perception and the behavior of consumers regarding products with an organic label, an experiment is chosen as the most suitable research method. When using an experiment, hypotheses can be tested easily by manipulating one independent variable. In the case of this research, the independent variable is the amount and nature of information on the product label. By manipulating the independent variable in a different way for each treatment, the effects on the respective dependent variables may differ in each case. The dependent variables in this research are respondents' behavior and perception of the labeled product. In the end, the differences in the dependent variables between each of the treatments can be attributed to the manipulations of the independent variables (Saunders et al., 2007).
3.2 Data Collection
was manipulated in different ways. Consequently, the effects of these different manipulations can be mapped in terms of respondents' differences in perception and behavior towards the different treatments. Veylinx offers the perfect platform to set up such experimental environments, and as such, the choice was made to use Veylinx for the experimental part of this research.
Veylinx
Veylinx is an online platform for academic research, specifically focused on the conduct of auctions. Veylinx uses the concept of so called Vickrey auctions. In Vickrey auctions, or second price auctions, the bidders are incentivized to bid their maximum willingness-to-pay. This means that for each bidder the highest expected outcome is achieved by following truthful bidding strategies (Vickrey, 1961). In this particular type of auction, the highest bidder gets a discount on his bid. In the case of Veylinx, as is illustrated by the name second price auction, the highest bidder pays the bid that was placed by the second-highest bidder. In this way, the bidders are incentivized to bid their maximum willingness-to-pay, since this is the price for which the bidder would be willing to attain the product and overbidding and underbidding may result in a net loss or not attaining the product. As such, it is thus the optimal strategy for each bidder to bid their maximum willingness-to-pay (Vickrey, 1961). Veylinx uses this concept of maximum willingness-to-pay in the following way. An invitation for an online auction is sent by email to the respondent panel. When respondents choose to participate in the auction, the link in the email leads them to a website where the auction “takes place.” Respondents get to see a form of advertisement and are asked to place their bid within a certain amount of time. Respondents are not able to see bids that are already placed, and are according to the theory on Vickrey auctions likely to bid their maximum willingness-to-pay. However, it could of course occur that respondents are not interested in the product, and as such they are also able to bid zero euros, since this is their maximum willingness-to-pay. After bidding, the respondents are led to a second webpage which offers the researcher the possibility to ask respondents a maximum of 5 predetermined control questions. These questions
can be used by the researcher for finding certain relationships and correlations or for use as control variables.
Next to the highly effective way of measuring respondents' willingness-to-pay and the possibility to ask respondents several follow up questions about the product directly after the auction, Veylinx is also a very suitable research platform because of the rather large amount of potential respondents. At the moment, around 2500 respondents are linked to Veylinx and the average response rate is around 30%. Moreover, each researcher is obliged to recruit between 50 and 100 new respondents for the platform, which ensures growth and possibly also qualitative improvement of the platform. Concerning the data gathered through the use of the follow up questions, it is worth noting that the data collected on respondents after each auction is saved, leading to an accumulation of data variables on all respondents. This data is in turn available for use by future researchers, leading to valuable insights, especially about respondents that are active users of the platform.
An important disadvantage connected to the way in which respondents are recruited for the platform arises because of the use of convenience sampling. This type of sampling weakens the external validity of the experiment, which has its implications regarding the generalizability of the results (Saunders et al., 2007).
Regarding this particular auction it is useful to get into the specific ways in which the experiment was set up and what the different manipulations of the respective treatments actually were. The advertisement for the product at auction, a 5 liter can of olive oil, was designed in 5 different ways. In this way, 5 different treatments can be used to unveil differences in respondents' behavior and perception of the advertisements.
3.3 Treatments
The five different treatments that make up this research can be viewed as 5 advertisements of a can of olive oil, which are all five the same, except for the point of manipulation. The point of
manipulation in this case is the amount and nature of information that is placed on the advertisement. In order to manipulate the different treatments in a relevant, yet consistent way, 4 units of information regarding organic product attributes and 1 unit regarding authentic product attributes were designed. The information units were designed in such a way that they provide clear information on the organic and authentic nature of the product. No above-average knowledge of either concept is needed in order to be able to understand the information on the advertisement. Moreover, since the auction is meant to imitate an actual consumer decision making process, the units of information were designed to be as short as possible without withholding relevant information. This is done because, in accordance with Information Theory (Nelson, 1970), when it takes a consumer too much time to evaluate a certain attribute, the relative cost will become too high, and the consumer will refrain from taking the attribute in question into consideration during the decision making process. The five different treatments can be found in Appendix Y.
Information Units
The five information units that were formulated were the following:
(1) This product completely follows the guidelines of the EU organic certificate. (2) Does not contain pesticides or other chemical detergents.
(3) Produced without the use of artificial manure or other unnatural growth accelerators. (4) Producers with the EU organic certificate are subject of control at least once a year. (5a) Produced on the hills of Troina, in Sicily using olives coming from trees that are over a century old.
(5b) Produced according to authentic production methods for generations and generations.
Treatment 1 consisted of just information unit one (1). This information unit was used to set up the
product was produced following its guidelines. In this treatment, there are not yet any cues that highlight the benefits of the actual organic nature of the product.
Treatment 2 consisted of the combination of information units one (1) and two (2). In extension to
the organic logo and information unit one (1) to help consumers identify the logo, information unit two (2) serves as an indication as to what one of the actual benefits of the organic nature of the product is. The lack of pesticides or other chemical substances may prove to be a sign for
consumers why the consumption of organic products is actually preferable over the consumption of regular products.
Treatment 3 consisted of the information units one (1), two (2) and three (3). This treatment
extended on the second treatment, with the information unit stating that there were no unnatural products used while producing the olive oil. Again here, the benefits that consumers may assume when reading this information, might help to make more clear the actual benefits of organic products.
Treatment 4 contained information units one (1), two (2), three (3) and four (4), adding to treatment
3 one extra unit of information with claims about the control program that goes with the EU organic certificate. This information unit may serve particularly to enhance the factor of trust that
consumers have towards the producer of the product and especially the institution that gives out the certificate in question.
Treatment 5 served as an alternative way to build trust between producer and consumer. In this
treatment information unit 1 was combined with information unit 5a and 5b, that both state claims on the authentic origins of the product. In this way, this type of information may serve as an alternative way to enhance trust and perception of organic nature of the product.
In the next paragraph, paragraph 4, the results of the analyses will be presented.
4. Results
In this section the results of the analyses will be reported. First some descriptive characteristics on the nature of the sample will be presented, after which descriptive statistics on some relevant variables will be shown. Then, some general correlations will be investigated. Finally the four hypotheses will be tested and to conclude some alternative findings will be reported.
4.1 Descriptive statistics.
4.1.1 Sample characteristics.
In this particular experiment 922 of the total of around 3200 respondents actually participated and placed a bid. This is a response rate of 28,8 %.
An initial quick scan of the data led to the identification of two strange values in terms of age. Two respondents indicated a birth year of 1900. This is an invalid value since the oldest person living in the Netherlands was born in 1902. A further look towards the characteristics of these cases led to the insight that they were “zero-bidders” and that they were missing a great amount of variables. The decision was made to exclude these cases.
After the exclusion of the two above mentioned cases, from the total of 920 respondents that remained, 507 were men (55,1 %), and thus 413 were female (44,9 %). When we compare these figures to the total population, which is the population of the Netherlands, these figures are
reasonably coherent. The population of the Netherlands consists for 49,5% of men and for 50,5% out of women.
When regarding the sample in terms of age, we see that 92,2 % of the sample is between 20 and 65 years old. The difference with the population arises especially because of the fact that the platform that Veylinx offers is recruited for by students. As such, especially the high difference in the
category 20-39 years old can be explained. Moreover, by using data collected in earlier auctions, the education level of 618 of the 920 respondents was determined. The fact that the sample has a very high level of education when compared with the average of the population, is also explained by the recruitment method that Veylinx uses.
Table 4.1.1.1: Average age and gender distribution Dutch population and sample
Gender: Men Women
Average age Dutch population 39.9 41.7
Average age sample (N=920) 41.1 41.9
Gender distribution Dutch population 49.5% 50.5%
Gender distribution sample 55.1% 44.9%
Table 4.1.1.2: Age distribution Dutch population compared to sample
Age: <20 20-39 40-64 65-79 >80
Dutch Population 23.1% 24.6% 35.5% 12.6% 4.2%
Sample (N=920) 1,5% 47,7% 45% 5,5% 0,3%
Table 4.1.1.3: Education Dutch population compared to sample Highest educational level: None or other Primary school Secondary school MBO HBO WO / University Dutch population 0,6% 8.4% 23% 40.1% 19.3% 9.2% Sample (N=618) 3,4% 0.3% 10,2% 23% 26,7% 36,4% Source: CBS, 2014
4.1.2 Descriptive statistics of the treatments and willingness to pay.
Now that we have taken a look at the characteristics of the sample, it is now time to look at some first descriptive statistics concerning the data. A first thing that is worth looking at is the difference in bid amounts per treatment. The 920 respondents were randomly assigned to one of the five treatments. This ended up in the following distribution of the respondents over the 5 treatments: Treatment 1; N= 165. Treatment 2; N= 180,Treatment 3; N= 184,Treatment 4; N= 190,Treatment 5; N= 201. Now since the number of respondents per treatment is relatively high, and at least much higher for every treatment than the minimum required N = 30, there is no reason to expect that the somewhat unequal distribution of respondents over the treatments will have any effect. In the following table the mean bids per treatment are presented.
Table 4.1.2.1: Descriptive Statistics of the different treatments.
Treatment N Mean Std Dev SE Mean Minimum Maximum
1 165 6,89 7,92 61,7 0 45 2 180 6,62 7,28 54,2 0 45 3 184 6,93 7,35 54,3 0 40 4 190 7,11 7,27 52,8 0 50 5 201 7,31 7,83 55,2 0 50 Total 920 6,98 7,52 24,8 0 50
When looking at these descriptive statistics, a first thing that is to be noticed is that the differences in means between the 5 treatments are quite small. The difference between the highest and lowest mean is just 69 cents. A second thing that hits the eye is that the standard deviations are high when regarding the means. This indicates that the data are spread out over a large range of values. This is also noticeable when looking at the height of the maximum bids in respect to the means. When looking at the minimum bids we notice that they are zero for every treatment. Now this is caused by the fact that respondents actually do not know in advance what type of product will be up for
well be that a large amount of respondents is for some reason not interested in the product, and thus places a zero bid. Obviously these zero bidders have a large impact on the distribution of the data.
Normality, kurtosis and skewness of the data.
When checking the nature of the data further, it is important to check whether the data are actually normally distributed, and what the kurtosis and skewness values of the data are. A look at the histogram of the frequency distributions of the willingness-to-pay shows that the data are not normally distributed (see appendix X, 4.1.2.A, figures 1, 2 &3). The histogram shows a cluster on and near the zero bid value, and moreover, throughout the distribution, high peaks can be observed. A quick look at these peaks gives the possible explanation that respondents are likely to place their bids in round numbers, or multiples of 5 euros. Analyzing the data further, we see that the data are positively skewed (skewness value = 1.757, SE = 0,081) and leptokurtic, (kurtosis value = 4.706, SE = .161) (see appendix X, 4.1.2.B), which confirms the high and sharp peaks that are observed when regarding the histogram. Regarding the skewness and kurtosis of the data, since we are not dealing with the entire population but just with a sample, z-scores can be computed as a check. The z-score for skewness in this case is 21,7, which is far above 2 and thus this confirms that there is a positive skew. The z-score for kurtosis is 29,2, which confirms that the data have positive excess kurtosis.
Another way to check the normality of the distribution of the data is to perform the Kolmogorov-Smirnov and the Shapiro-Wilk tests. These tests confirm that the data are not normally distributed with values of 0.178 (p = 0,000) for the Kolmogorov-Smirnov test and 0.829 (p=0,000) for the Shapiro-Wilk test.
As was stated above, the relevance and influence of zero bidders should be kept in mind. An alternative way to look at the collected data is to exclude these zero bidder cases. In the following table, the descriptive statistics for the different treatments are displayed again, however this time
excluding the zero bids.
Table 4.1.2.2: Descriptive Statistics of the different treatments – excluding zero bids.
Treatment N Mean Std Deviation SE Mean Minimum Maximum
1 125 9,09 7,92 70,87 0,50 45 2 138 8,64 7,18 61,16 0,01 45 3 143 8,92 7,21 60,30 1,00 40 4 156 8,66 7,14 57,15 0,01 50 5 154 9,54 7,67 61,78 0,10 50 Total 716 8,97 7,41 27,68 0,01 50
After excluding the zero bids from the analysis of the mean per different treatment, it is no surprise that the means of each of the treatments have gone up. The differences between the treatments are still small though and they are also not ascending or descending in a way that is coherent with the overall assumption that more information on the product label would lead to higher willingness to pay for the product. Something that does look interesting is that the mean for treatment five, which was the treatment emphasizing the authenticity of the product, is the highest. However, since the results are not at all significant (p = 0,93), there is not really anything that can be said based on these simple primary analyses.
Normality, kurtosis and skewness of the data.
Again here, the data should be checked in terms of normality distribution, skewness and kurtosis. The data are still positively skewed (skewness value = 1.781, SE = 0,091) and remains leptokurtic, (kurtosis value = 4.968, SE = .182) (see appendix X, 4.1.2.B), with high and sharp peaks still occurring in the frequency distribution histogram. The relative z-scores are 19,6 for skewness and 27,3 for kurtosis.
The Kolmogorov-Smirnov and the Shapiro-Wilk tests confirm the non-normality of the distribution again with values of 0.158 (p = 0,000) for the Kolmogorov-Smirnov test and 0.850 (p=0,000) for
