The influence of transparency on implicit and explicit attitudes.
Master thesis
Submitted by:
Annemiek Scheltus
Student nr:
10084754
Supervisor:
J. Demmers
Second supervisor:
To be announced
MSc Business Studies – Marketing track
University of Amsterdam – Faculty of Economics and Business
Final 30/6/2014
2 TABLE OF CONTENTS Chapter Page(s) Acknowledgement ……….…….. Abstract ……….……….. Introduction ……….………. 1.1 Problem statement ……… 1.2 Aims and objectives of the study ………. 1.3 Thesis structure ……….… Theoretical framework ……….…….... 2.1 Transparency ……….…………... 2.2 Attitude ……….……….. 2.3 Purchase behavior ……….……….. 2.4 Conceptual framework ……….….. Methodology ……….……… 3.1 Research design ……….…………. 3.2 Pre-test ……….………... 3.3 Implicit Association Test ………... 3.3 Task procedure ……….………….. 3.4 Task 1 ……….……… 3.5 Task 2 ……….……… Results ……….………. 4.1 Descriptive results ……….………. 4.2 Implicit results ……….…………... 4.3 Explicit results ……….…………... 4.4 Purchase results ……….…………. 4.5 Correlation between IAT and explicit findings ……….. Conclusion ……….………... 5.1 Theoretical contributions and managerial implications ………. 5.2 Limitations and suggestions for future research ……… References ……….………... Appendices ……….……….. 4 5 6 8 9 11 11 11 15 19 21 22 22 23 26 28 29 31 32 32 33 37 37 39 40 41 43 46 51
3 LIST OF FIGURES AND TABLES
Figure/table Page(s)
Figure 1: Dimensions of transparency
Figure 2: Typology of transparency for marketing management research Figure 3: Conceptual model
Figure 4: Pure Tea and Orange Delight Figure 5: Start of the IAT
Figure 6: Graph of D-score Figure 7: Histogram of D-score
Table 1a: Comparison in group number between the three conditions Table 1b: Comparison in age and gender between the three conditions Table 1c: Comparison in educational level between the three conditions Table 2: IAT D- score procedure
Table 3: D- score of respondents
Table 4: Explicit preference of respondents
Table 5: Descriptive statistics of valuation statements
Table 6: Correlation analysis between D-score and preference
12 13 21 24 30 35 36 33 33 33 34 35 37 38 39
4 Acknowledgement
I would like to thank my supervisor J. Demmers for his guidance, support, advice and comments on this thesis. His support has been of great value in writing this thesis. Furthermore, I would like to thank all respondents for taking time to help me with my research. Finally, great words of appreciation directed towards my fellow students, family and friends for their support and positive motivation.
5 Abstract
Transparency is becoming increasingly important for consumers. Knowing information decreases uncertainty and leads to well-thought decisions. However, consumers only want to know information that supports their beliefs and which gives a positive feeling. As a result, consumers try to avoid information that can lead to an undesired action or a change of beliefs. This study examines this difference with implicit and explicit attitudes. Implicit attitude is measured using an implicit association test and explicit attitude is measured using a self-report questionnaire. Subjects explicitly preferred the transparent marketed product over the non-transparent product, whereas subjects implicitly preferred the non-transparent marketed product over the transparent product. Further, this study reports that purchase intentions are stronger related to explicit attitude than to implicit attitude. Finally, no correlation was measured between explicit and implicit findings.
Keywords: transparency ▪ implicit attitude ▪ explicit attitude ▪ IAT ▪ purchase behavior ▪
6 1. Introduction
A third of the United Kingdom’s (UK) food products are not what they say they are or have misleading ingredients lists (NHD, 2014). This is shown by the Food Standards Agency’s (FSA) research. The FSA checked hundreds of food samples, which were taken in West Yorkshire (Guardian, 2014). Results showed mozzarella that was less than half real cheese, ham on pizzas that was poultry, fruit juice with brominated vegetable oil, which is designed for use in flame retardants, and vodka that had not be made from alcohol derived from agricultural produce, but from isopropanol which is used in antifreeze and as an industrial solvent. The Dutch food agency (NVWA) does not want to address whether this is the same in the Netherlands, but Hilde de Vries, the director of Foodwatch, stated that she did not see a reason why it should be different in the Netherlands (NHD, 2014). A recent Dutch broadcast showed that the brown granules in Lipton Green Tea and Yellow Label Tea largely existed from sugar, but sugar was not listed as an ingredient (KRO, 2014).
Dutch law states that the name of a product or the image on a product’s package does not need to be the main ingredient of the product (NVWA, 2014). This is allowed as long as the percentages of each ingredient are mentioned in the ingredients list. Therefore, it is possible for companies to mislead the consumer with good-looking packaging and slogans. However, the real product information can differ from the package that is presented to consumers.
Since the 1950s, the consumption of products and services started to grow. Consumers were thought to act rationally, and according to the neoclassical economics theory, consumers should only buy products or services that maximize their satisfaction or if the functional benefits outweigh the costs (Baines, Fill, & Page, 2011). Over the years, consumer were becoming better informed. Television and radio were an important part of this change. In the Netherlands, for example, a 1965 a television program called Koning Klant (King Consumer)
7
tried to make consumers more conscious about what they bought and empowered them against all kinds of sales techniques (VARA, 2014).
Recent research of IBM (2010) showed that current consumers are smarter, more diverse, and more demanding because of the internet, social media, and a multitude of consumer programs. Consumers can now access much information about companies more easily than in the past. Consumers use this information to shop more carefully and share this information with others. According to IBM, a third of the consumers followed a company on a social networking site. Because consumers are better informed, the behavior of consumers towards products and services is changing.
Previous research of human behavior assumed that the roots of human behavior are in the higher order processes of deliberate reasoning. Presently, human behavior is viewed as a result of automatic processes that may occur spontaneously and outside of a person’s awareness or control (Bargh, 1997; Moors & De Houwer, 2006). Contemporary research on attitudes have often contrasted explicit, conscious attitudes with implicit, unconscious attitudes (Gawronski & Bodenhausen, 2006). Although both explicit and implicit attitudes derive from previous experiences, several researchers argued that implicit and explicit attitudes are two conceptually distinct cognitive processes (Devine, 1989; Greenwald & Banaji, 1995).
Research has shown that consumers explicitly say they prefer accurate and realistic information rather than attractive, inaccurate information (Cohn & Wolfde 2012, 2013; Bhaduri & Ha-Brookshire, 2011). At the same time, consumers show different behavior by avoiding information that creates knowledge (Sweeny, Melnyk, Miller, & Shepperd, 2010), purchasing products that only reflect their values, or using products and brands to create and represent desired self-images to others or even to themselves (Escalas, 2004).
8 1.1 Problem statement
According to a survey from Cohn and Wolfe (2013), transparency has become more important for consumers. The results showed that 47% of the consumers in the UK would be extremely angry if a company failed to report some of the ingredients in their food and 53% considered transparency and honesty as important factors in the decision making process (Cohn & Wolfe, 2012, 2013). If a company is transparent and also exposes negative information or consequences, the framing theory states that, as a result, consumers will be more critical towards the product or service and more willing to choose the non-transparent product (Kahneman & Tversky, 1984; Meyerowitz & Chaiken, 1987).
On the other hand, the research of Escalas (2004) showed that there is a positive relationship between narratives and self-brand connections, which in turn has a positive relationship with brand attitudes and behavioral intentions. A story may lead to a self-brand connection if the story creates the belief for the consumer that the brand contributes to his or her self-related, psychological needs. The connection between a consumer and a brand can be enhanced by advertisement. Ads are able to involve and entertain consumers to create favorable brand associations. Showing a favorable, good-looking ad makes consumers focus more on the attractiveness and likeability of that ad, instead of brand attributes or possible criticisms (Baumeister & Newman, 1994).
In theory, a company can have a good product and be transparent, but reality shows differently. Most market offerings vary with the level of transparency (Hsieh, Chiu, & Chiang, 2005; Weathers, Sharma, & Wood, 2007; Wright & Lynch, 1995). Not being transparent can help companies develop a competitive advantage because of lower information acquisition costs. Many companies try to mislead the consumer with positive and attractive information, so that transparency becomes less important. Misleading the consumer
9
means that there is a certain discrepancy between what the ad shows and the actual product performance (Russo, Metcalf, & Stephens, 1987).
Studying transparency is also an interesting topic because 83% of global consumers think companies today are very clever in how they try to mislead or hide information from the public (Cohn & Wolfe, 2013). On one hand, consumers want to know everything, but on the other, they do not trust the information that is being presented.
Moreover, several attitude models do not distinguish between explicit and implicit attitudes, but treat attitudes as a unitary construct (Kruglanski & Thompson, 1999; Petty & Cacioppo, 1986; Van Overwalle & Siebler, 2005). As such, these models leave an explanatory gap for any differences that may emerge between implicit and explicit attitudes. Consumers deliberately (explicitly) say that they want to know the truth (Cohn & Wolfe, 2013), but their automatic (implicit) behavior shows something different because consumers buy the products that look attractive, have favorable slogans (e.g., 0% fat), and gives them good feelings. According to Messner and Vosgerau (2010), implicit attitudes may be a better predictor of behavior than explicit attitudes. If correct, this would have managerial and theoretical implications for marketers.
1.2 Aims and objectives of the study
This research will examine the influence of transparency on implicit attitudes, explicit attitudes, and purchase behavior. The study will explain why consumers say that they want the truth behind a product but show different behavior by buying products that are not transparent. Furthermore, there is still little known about transparency and the way companies should handle it. Due to these contradicting results, the purpose of this study is to better understand how consumers perceive transparency. By focusing on two constructs of transparency, amount of information and reliability of information (explained in the next
10
chapter), this research will try to understand the relationship between transparency and implicit and explicit attitudes.
Therefore, the research question is as follows: What is the relationship between the
amount of information and the reliability of that information on consumers’ implicit and explicit attitudes and the influence on their purchase behavior?
In this study, a two-way experiment will be conducted in which the extent of the two transparency constructs’ influence on explicit and implicit attitudes will be investigated, as well as whether these results match each other. To measure implicit attitude, an implicit association test (IAT) will be used. Explicit attitude is measured through self-report questionnaires.
1.3 Thesis structure
This study is constructed by using primary research and is structured in this way: Chapter 2 discusses previous and current literature that have been gathered from several academic journals and books and provides hypotheses and the conceptual model; Chapter 3 focuses on methodology and research design; Chapter 4 examines and evaluates the results from the research and analyzes the findings; and in Chapter 5, the theoretical and managerial contributions are explained, followed by limitations and suggestions for further research.
11 2. Theoretical framework
In this section, the existing literature related to the research problem will be addressed in order to define the research topic more precisely. The following concepts are discussed: transparency, attitude, and purchase intention.
2.1 Transparency
Today’s consumers are more conscious about their environment and society. They demand transparency and sustainable products due to increased awareness of their environment and advanced technology in communication (Bhaduri & Ha-Brookshire, 2011). Because of better access to information, there are more educated consumers with higher information needs (Fournier & Avery, 2011). Transparency is defined as “visibility and accessibility of information especially concerning business practices” (Merriam-Webster, 2010), and sustainability can be defined as “developments that meet present needs without compromising the ability of future generations to meet their needs” (World Commission on
Environment and Development, 1987, p. 8).
Research showed that transparency is becoming increasingly important for consumers. In 2012, 53% of consumers considered transparency and honesty important factors in their decision making processes (Cohn & Wolfe, 2012). In 2013, this had grown to 66% (Cohn & Wolfe, 2013). After quality (89%) and price (84%), transparency is the most important factor in the decision making process (Cohn & Wolfe, 2013). Because consumers have more and better information, they can make well-thought out decisions to better avoid uncertainty.
There is already a large body of scholarly work on marketing practice and marketing management. However, only a few works addressed transparency in those contexts (Eggert & Helm, 2003). Previous research about transparency has dealt with financial markets, transaction cost economics, and value transparency (Hultman & Axelsson, 2006; Lamming, Philips, & Caldwell, 2002; Visnawath & Kaufmann, 2001). In financial markets, research
12
showed that greater availability of reliable and timely information improves efficiency and resource allocation and increases the prospects for growth (Visnawath & Kaufmann, 2001). In transaction cost economics, scholars have researched how cost and market imperfections may be dealt with (Williamson, 1981), and within the area of industrial marketing, the concept of transparency has been elaborated on in terms of value transparency and cost transparency. In all cases, the key characteristics behind transparency are “to see through” something and that information is being shared that is not usually shared. A lack of transparency may exist if there is no access to information or if information is irrelevant, inaccurate, or misrepresented.
Transparency dimensions
Although the concept of transparency is stated in several studies (e.g., Bhaduri & Ha-Brookshire, 2011; Cohn & Wolfe, 2012, 201; Hultmann & Axelsson, 2007), there is still the need for a clear definition in the literature. The core concepts of transparency used by several researchers are access, comprehensiveness, relevance, quality, and reliability (see figure 1) (Bhaduri & Ha-Brookshire, 2011; Visnawath & Kaufmann, 2001)
Figure 1: Dimensions of transparency
The first dimension (access) states that information should be available to all. There is a need for timely, equitable dissemination of information. This access should be provided by newspapers, television, radio, internet, public notes, and word of mouth. Lack of accessibility is detrimental to transparency because it limits the opportunity to interpret and respond to
Transparency
Access Relevance Comprehensive Reliability and quality
13
information. The second dimension (relevance) states that information that is shared should be relevant to the user. The third dimension (comprehensive) states that the information that is shared should be understandable, presented in clear and simple terms. Lastly, the fourth dimension (reliability and quality) states that to be effective, information should be fair, timely, consistent, complete, and reliable. Additionally, maintaining a transparent supply chain is important for building brand image and brand loyalty (Strutnin, 2008). Furthermore, Carter and Rogers (2008) suggested that to be transparent within business operations built reputation and maintained legitimacy.
Taking into consideration all of the above, transparency exists when there is information shared and that information can be trusted and thus is reliable. In this study, therefore, the amount of information and reliability of information will be the core concepts driving transparency.
Types of transparency
There are several types of transparency. So far, transparency has been mainly discussed regarding how a business can be transparent and how to use transparency to increase its sales (e.g. Cohn & Wolfe, 2012, 2013). However, there are many more types of transparency. First of all, organizational transparency is the availability of firm-specific information to those outside publicly traded firms (Bushman, Piotroski & Smith, 2004). Hultmann and Axelsson (2007) have defined four more types within marketing management and extended these by adding three facets (see figure 2).
Previous literature on marketing management had a strong focus on buyer-supplier relationships (Dwyer, Schurr & Oh, 1987; Ford, 1980) and on relationship marketing (Blois, 1994). Trust and commitment were key to these relationships. Trust has been an important theme in previous studies on transparency due to the interrelation between trust and sharing
14
information. Having trust in a business partner increases the likelihood of sharing information (Morgan & Hunt, 1994).
Figure 2: Typology of transparency for marketing management research
The four types of transparency are technological, organizational, supply, and cost/price transparency. The three added facets are degree of transparency, direction of transparency, and distribution of transparency. Degree of transparency depends on the extent to which a company is being completely transparent, being translucent in some aspects, or is not being transparent at all (Lamming, Caldwell, Harrison & Philips, 2001). Direction of transparency depends on if the information flow is one-way or two-way (Harrington, 1995), and distribution of transparency depends on if the relationship is direct or indirect (Eggert & Helm, 2003).
Some authors distinguish between horizontal and vertical transparency (Grimmelikhuijsen, 2007; Heald, 2006). Horizontal transparency takes place between the environment and the organization. There are two types of horizontal transparency: inward and outward. Inward means that a consumer from outside the organization can see what happens inside the organization. Conversely, outward means that someone within the organization can see what happens outside the organization. Vertical transparency describes the hierarchical relationship in an organization. There are again two types of vertical transparency: top-down and bottom-up. Top-down transparency means that a manager can monitor the behavior of the
15
employees, and bottom-up means that the employees can monitor the activities of the manager (Grimmelikhuijsen, 2007; Heald, 2006).
In this study, the degree of transparency is manipulated to determine if high versus low transparency influences a consumer’s implicit and explicit attitudes. A high transparency story will score high on the transparency dimensions, and a low transparency story will score low on the transparency dimensions. The focus of transparency type lies on inward horizontal transparency, meaning that a consumer from outside the organization can see what happens inside the organization.
2.2 Attitude
Over the past decades, considerable research focused on how attitudes are formed and changed (Ajzen & Fishbein, 1980, Kassarjian & Kassarjian, 1979; Petty & Cacioppo, 1983). Attitude can be described as “a learned predisposition to respond in a consistently favorable or unfavorable manner with respect to a given object” (Fishbein & Ajzen, 1975, p. 6).
Implicit and explicit attitudes
Numerous research studies have focused on implicit social cognition (Roediger, 1990a; Fazio, 1990; Greenwald & Banaji, 1995). Implicit social cognition suggests that behavior and social judgment are the result of an individual’s implicit and explicit attitudes. As mentioned before, implicit attitudes are automatic evaluative dispositions that typically occur without conscious reflection (Fazio, 1990). In contrast, explicit attitudes are reactions formed through a cognitive, controlled process (Wittenbrink & Schwarz, 2007). Although both implicit and explicit attitudes derive from previous experiences, some researchers argue that they are two distinctive processes (Devine, 1989; Greenwald & Banaji, 1995). Implicit attitudes might be a better predictor of behavior than explicit attitudes because consumers might want to hide their negative attitude and give socially desirable answers.
16
Previous research on implicit and explicit attitudes have mainly focused on racial attitude and language (Pantos & Perkins, 2012). Research on attitude change mostly does not distinguish between implicit and explicit attitudes, instead treating attitude as a unitary construct (Kruglanski & Thompson, 1999; Petty & Cacioppo, 1986; Van Overwalle & Siebler, 2005). Because implicit attitude might be a better predictor of behavior (Messner & Vosgerau, 2010), this study will measure both implicit and explicit attitudes.
APE model
Gawronski & Bodenhausen’s (2006) associative propositional evaluation model (APE) is a dynamic cognitive model that explains how attitudes are formed through two processes: associative (automatic) and propositional (thoughtful) processing. The implicit-explicit distinction is interpreted within the APE framework. Associative processes are characterized as affective, immediate reactions towards an object, resulting from unconscious retrieval of previous existing information in memory. In contrast, propositional processes try to determine the validity of evaluations and beliefs by assessing their consistency with other propositions that are relevant (Gawronski & Bodenhausen, 2006; Pantos & Perkins, 2012).
In summary, an attitude is a response to an object or situation. It can be formed through two processes: associative and propositional processing. There are two types of attitudes: implicit and explicit.
When the influence on implicit or explicit attitude is stronger, this can be explained using the elaboration likelihood model (Petty, Cacioppo, & Schumann, 1986). Depending on the level of involvement, the influence of an advertisement will be higher for explicit attitude or implicit attitude. If a consumer is highly involved, the central route to persuasion is most effective (Petty et al., 1986). Relevance of the presented information or the argument quality are the factors that are most important for a positive attitude change. If the information is weak or specious, unfavorable attitude change will result. This attitude change is a cognitive
17
process and, therefore, will have a higher influence on explicit attitude. This is because explicit attitude is formed through a cognitive, controlled process (Wittenbrink & Schwarz, 2007). If a consumer is less involved, the peripheral route to persuasion is most effective (Petty et al., 1986). The attractiveness of ad cues, humor, or presentation by a famous endorser are factors that influence attitude. Consumers that have an attitude change when following the peripheral route are temporarily non-predictive of behavior (Petty et al., 1986). This process will have a higher influence on implicit attitudes because implicit attitudes are also automatic, immediate reactions and not based on cognitive processes (Fazio, 1990). Thus, under high involvement, attitude will be more explicitly influenced and under low involvement, attitude will be more implicitly influenced.
Information avoidance
Although knowledge is valuable and important to form and change attitudes, people do not always seek it and sometimes try to avoid it. This information avoidance can explain why people continue to buy products that make them feel and look good instead of buying products that reveal both positive and negative information. Information avoidance can be defined as any behavior intended to delay or prevent the acquisition of available, but potentially unwanted, information (Sweeny, et al., 2010). Previous research about information avoidance is mainly focused on selective exposure. Selective exposure is grounded in dissonance theory and states that people only seek information they want to find (Smith, Fabrigar, & Norris, 2008). People choose to avoid information if it may demand an undesired action or change of beliefs or if the information itself may cause unpleasant emotions or diminish pleasant emotions. Therefore, consumers only seek information that supports their beliefs; that brings them happiness, pride, or relief; that reduces negative feelings of fear or worry; or that provides a sense of closure to uncertain situations (Meissen, Mastromauro,
18
Kiely, McNamara, & Myers, 1991). If consumers have doubts about the values that are presented by the company, half of those consumers would stop buying the product, nearly a third would encourage friends to stop buying it, and a quarter would even support a boycott of the company (Cohn & Wolfe, 2013). This change of beliefs and the resulting actions are undesired (Sweeny et al., 2010). Therefore, consumers would try to avoid this situation. Instead, consumers seek information that brings them happiness and supports their beliefs. A positive brand story can help consumers achieve a positive feeling and does not require a negative change of beliefs or an undesired action.
Brand stories can help create a link between the brand and the consumer’s self-concept (Escalas, 2004). Consumers use brands and products to create desired self-images and present these images to others. Therefore, the products they buy should make them look good. This supports the view that consumers try to avoid information that demands a change in beliefs but that consumers do prefer a positive brand story because it supports their beliefs or brings them happiness.
A company can have a good brand story and an honest product that supports the belief of consumers. However, several studies showed that market offerings and communication practices vary with the level of information asymmetry (Wright & Lynch, 1995; Hsieh et al., 2005; Weathers, et al., 2007). High levels of information asymmetry can help companies to develop a competitive advantage through lower information acquisition costs. If elements of the brand story convey information consumers want to hear, the company positively influences the readers’ attitudes. This might be a reason why some companies are not transparent.
Consumers are more conscious about products and companies because of better access to information (Bhaduri & Ha-Brookshire, 2011). A large number of consumers see transparency as one of the most important factors in the decision making process (Cohn &
19
Wolfe, 2013). Because of better information, consumers can make well-thought out decisions and avoid uncertainty (Fournier & Avery, 2011). Therefore, it is expected that consumers explicitly prefer transparency. However, possessing information can also create negative consequences, changes of beliefs, or undesired actions. Because of this, consumers do not always seek information and may even avoid it. Consumers prefer information that supports their beliefs or brings them happiness (Meissen et al., 1995). A positive brand story can, therefore, positively influence a consumer’s implicit attitude. Therefore, the following hypotheses are proposed:
H1: Participants implicitly prefer a non-transparently marketed product over a transparently marketed product.
H2: Participants explicitly prefer a transparently marketed product over a non-transparently marketed product.
2.3 Purchase intentions
Several studies measured the impact of corporate behavior and credibility on attitude and purchase intentions (Creyer & Ross, 1996; Lafferty & Goldsmith, 1999). Results showed that consumers were more likely to purchase from brands with ethical, credible business practices than those that had unethical, untrustworthy practices. In this study, consumer purchase intention was defined as “the buyer’s self-instruction to purchase the brand” (Rossiter & Percy, 1998, p. 126). According to the theory of reasoned action, a person’s intention to show certain behavior is a function of his or her attitude and subjective norms (Ajzen & Fishbein, 1980). Attitude toward a behavior is personal in nature and is a function of behavioral beliefs. This means that a person who believes that behavior will lead to a favorable outcome will be more willing to perform that behavior than one who does not.
20
Ha-Brookshire and Hodges (2009) proposed that the theory of reasoned action can be integrated within a consumer’s perceived value. Holbrook (1999, p.8) noted that a consumer’s perceived value is an experience that results from the consumption of a good or service. There can be two types of benefits after an experience: hedonic or utilitarian (Chitturi, Raghunathan & Mahajan, 2008). Hedonic values are psychological values and derive from a pleasurable, enjoyment-related experience, while utilitarian values derive from a functional, practical, and task-oriented experience (Carpenter, Moore, & Fairhust, 2005; Chitturi et al., 2008). Ha-Brookshire and Hodges (2009) tried to explain the mix in the context of used clothing donation behavior, where the hedonic (feeling better by donating) and utilitarian (clean and empty closet) values both contribute to the need of consumers to be socially and environmentally conscious. Transparency is essential in evaluating business practice, and it is also sought by today’s consumers (Bhaduri & Ha- Brookshire, 2011). Thus, both consumer perceived value and the theory of reasoned action might be useful in understanding the relationship between transparency and purchase intentions.
Consumers are more likely to purchase from brands with transparent business practices (Creyer & Ross, 1996; Lafferty & Goldsmith, 1999), and transparency is currently an important factor in the decision making process (Cohn & Wolfe, 2013). Therefore, it is expected that consumers’ purchase intentions are strongly influenced by the presence of transparency. Purchase intention is a function of a consumer’s attitude (Azjen & Fishbein, 1980). Because explicit attitude is a process of “evaluative judgments that are based on syllogistic inferences derived from any kind of propositional information that is considered relevant for a given judgment” (Gawronski & Bodenhausen, 2006, p. 694), it is expected that purchase intentions will be influenced by the outcome of explicit attitudes. Therefore, the following hypothesis is stated:
21 H3: Purchase intentions are more strongly related to explicit attitudes than to implicit attitudes.
2.4 Conceptual Framework
22 3. Methodology
In the previous chapter, all relevant literature related to the research problem was discussed, the hypotheses were stated, and a conceptual model was presented. This chapter will give an overview of the research design, pre-test and task procedure, and the conceptual model’s design.
3.1 Research design
In this study, explanatory research will be conducted in order to find an answer to the research question. Explanatory research is defined as “Research that focuses on studying a situation or problem in order to explain the relationships between variables” (Saunders, Lewis & Thornhill., 2009, p.591). This definition is in line with the research purpose, namely to understand the relationship between the amount of information and the reliability of that information on a person’s implicit and explicit attitudes.
There are many ways to conduct empirical research: experiment, survey, case study, and so on. However, because of the explanatory purpose of this study and in order to understand the causal links between the variables, an experiment will be used. An experiment is generally associated with a deductive research approach (Saunders et al., 2009). A characteristic of the deductive research approach is the collection of quantitative data. After studying the present literature, hypotheses were developed and a research strategy was designed to test the different hypotheses (Saunders et al., 2009, p. 117). In order to test these hypotheses, a large sample of people will be needed to provide reliable results (Saunders et al., 2009, p.144). Questions and statements must be consistent and standardized in order to compare the answers of different people (Saunders et al., 2009, p.373).
The second task of the experiment will be self-administered. Self-administration saves time when collecting data and will guarantee anonymity, since there is no interviewer who has
23
to write down the respondents’ answers. This will reduce the chance of participant or subject bias and, therefore, will improve the reliability of the data (Saunders et al., 2009, p.156). Using a questionnaire will give more control over the research process and allow the collection of a large amount of data in a highly economic way (Saunders et al., 2009, p. 144). A disadvantage of a questionnaire is that the number of questions is limited; otherwise, the respondents could not finish the questionnaires.
This study will use a two-way experimental design. The independent variables will be the amount of information and the reliability of information. Implicit attitude, explicit attitude, and purchase behavior will be the dependent variables. Amount of information will be manipulated by the number of words. In the transparent condition, considerable information will be presented about a company, whereas in the non-transparent condition, little information will be presented. Reliability will be manipulated in the content of the information. In the transparent condition, information will be presented as relevant and reliable, confirmed with a quality mark, whereas in the non-transparent condition, the information will be presented as a nice story, but will not appear legitimate.
3.2 Pre-test
Before the main test could be conducted, a pre-test was needed to check if the chosen stories were indeed evaluated as transparent or non-transparent. The method used for this pre-test was an internet-mediated method. The website www.qualtrics.com was used to develop the questionnaire because it provides many options for different types of questions and is easy for respondents to use. Additionally, it saved time, since the data did not had to be entered manually (Saunders et al., 2009, p.365).
Prior knowledge about transparency efforts of a business influenced the intention of consumers to buy a product (Bhaduri & Ha-Brookshire, 2011). If the transparency effort was
24
familiar to a consumer, or if they were aware of it, purchase intention was stronger than when they did not know about the transparency effort. Therefore, fictitious brands were used in this study to overcome the chance that results were biased by prior knowledge.
Two convenience products were chosen: tea and orange juice. Two fictitious brands were designed: Pure Tea and Orange Delight (see figure 4). For each brand, a transparent story and a non-transparent story was written. The transparent story included the different dimensions of Visnawath and Kaufmann (2001), while the non-transparent story did not. All stories were measured on positivity because being transparent should not mean that the information that is revealed is seen as negative. For the implicit association test, different words are necessary as stimuli in the test. Therefore, the pre-test started with these two questions: Would you please write down all associations (thoughts) you have regarding tea,
(or respectively orange juice)? Subsequently, the preferences of the participants were
measured. This ensured that participants did not specifically prefer one product, so that the final results were more reliable.
25
After the associations and preferences were noted, participants were presented with four different stories: two about Pure Tea and two about Orange Delight. Each product had a transparent story and a non-transparent story. The goal was to test the four stories against the different dimensions of transparency. It was expected that the transparent stories would score higher on the transparency dimensions than the non-transparent stories. The questionnaire used can be found in Appendix A.
Sample
The total number of respondents who started the questionnaire was 36, however 7 questionnaires were not completed and, therefore, excluded from the sample. As a result, the final sample consisted of 29 respondents, of which 10 were men and 19 female.
Results
First, the tea and orange juice associations listed by the respondents can be found in Appendix B. The most frequently named words were noted for the IAT. The respondents’ preferences were as follows: 7 respondents (24%) had a preference for orange juice, 9 respondents (31%) for tea, and 13 respondents (45%) had no a preference. These results indicated that among the respondents there was no specific preference, thus the two chosen convenience products were appropriate to use in the main test.
Second, a reliability analysis was performed to check if the four transparency statements could form a scale. The analysis exposed a Cronbach’s alpha of 0.876, high enough to be statistically allowed to form a scale. In order to test whether there was a significant difference in transparency between means among the four different stories, a one-way analysis of the variance-test was used. This analysis exposed a highly significant difference between the four stories (F(3,112) = 78,505; p = 0.000). Story 1 (M = 4.01, SD =
26
0.41) and 4 (M = 4.18, SD = 0.37) were the transparent stories, whereas story 2 (M = 2.47, SD = 0.71) and 3 (M = 2.50, SD = 0.68) where the non-transparent stories. A contrast test using coefficients indicated that there was a significant difference between the two transparent stories and the two non-transparent stories (p = 0.000); no difference in mean was found between the two transparent stories (p = 0.251) nor between the two non-transparent stories (p = 0.863).
Another one-way analysis of variance was used to test whether there was a significant difference in how positively the stories were evaluated. The analysis exposed a significant difference between the stories (F(3,111) = 6,698; p = 0.000). A contrast test using coefficients indicated that there was a difference in positivity evaluation between the two transparent stories and the two non-transparent stories (p = 0.000). The two transparent stories, story 1 (M = 4.10, SD = 0.55) and story 4 (M = 4.07, SD = 0.53), were seen as more positive than the two non-transparent stories, story 2 (M = 3.51, SD = 0.82) and story 3 (M = 3.48, SD = 0.82). This corresponds with what this study expected, namely that consumers explicitly prefer to know the truth about products and these products are, therefore, more highly evaluated. No difference was found in positivity between the two transparent stories themselves (p = 0.864) or between the non-transparent stories (p = 0.852).
3.3 Implicit Association Test
Implicit attitudes are inaccessible to the individual because they are automatic, immediate reactions (Wittenbrink & Schwarz, 2007). Because of this, implicit attitudes are difficult to assess. Implicit attitudes can be captured only through indirect methods that do not rely on personal introspection. A method that can be used is the implicit association test (IAT) (Greenwald & Banaji, 2005). In the IAT, participants are presented stimuli sequentially on a computer screen and then have to sort them into categories. Response latencies reveal the
27
respondents’ automatic evaluations toward one category. Previous research has shown that the IAT is a valid measure of implicit brand attitudes (Maison, Greenwald & Bruin, 2004).
The difficulty of switching from one category rule to the opposite rule can create cognitive inertia (Messner & Vosgerau, 2010). The IAT effect depends on the order of the two IAT blocks. Cognitive inertia distorts individual IAT scores and diminishes the correlations between predictor variables and IAT scores when the block is being counterbalanced between subjects. Specifically, IAT effects are stronger when the compatible block precedes the incompatible block (Greenwald, Nosek & Banaji, 2003). The architect of the IAT, Anthony Greenwald, calls this order effect “the most noticeable internal validity of the IAT” (Greenwald & Nosek, 2001, p. 87). To overcome biased results due to cognitive inertia, half of the respondents proceeded through the block sequences 1-2-3-4-5-6-7 and the other half proceed through 1-5-6-7-2-3-4.
In addition to implicit attitudes, this study addresses explicit attitudes as well. Previous studies already examined whether the IAT related to explicit measures of prejudice (Greenwald & Banaji, 1995; Greenwald et al, 1998). With respect to the IAT, Greenwald et al. (1998) did not find any correlation between explicit measures and the IAT. Other studies (Dovidio, Kawakami, Johnson, Johnson, & Howard, 1997, experiment 2; Wittenbrink, Judd, & Park, 1997) did find relationships between explicit measures and the IAT.
This contradiction in results allows an opportunity to determine if this study will find a relationship between explicit and implicit attitudes. Because the IAT is a valid measure of implicit brand attitude, the IAT is chosen as the method to measure implicit attitudes in this study.
In this study, explicit attitudes will be measured through asking the preference of the respondent. Purchase behavior will be measured on the ranking of six different statements. Respondents have to rank statements from most important to least important when making a
28
product choice. Three of these statements are transparency statements and the other three statements are the opposite of transparency (Visnawath & Kaufmann, 2001). With these rankings, this study attempts to measure what people find more important when buying a product.
3.4 Task procedure
Previous research has shown that the order of testing has no effect on the outcome (Lane, Banaji, Nosek, & Greenwald, 2007; Nosek, Greenwald, & Banaji, 2005). In this study, the IAT was conducted before the questionnaire. There were two different tasks. In total, the tasks took 15 minutes to complete: approximately ten minutes for task 1 (IAT) and five minutes for task 2 (self-report questionnaire). Demographic questions were included in task 2. In advance of the IAT, respondents had to read two product descriptions, one about Orange Delight and one about Pure Tea.
Respondents were randomly designated to one of the three conditions. In condition one, the product Orange Delight was presented as a transparent product and Pure Tea as a non-transparent product. In condition two, the product Pure Tea was presented as a transparent product and Orange Delight as a non-transparent product. In condition three, the control condition, both products were presented as non-transparent. The different stories can be found in Appendix C.
The entire study was computer based. Each participant completed the IAT and the questionnaire on the researcher’s laptop. All participants completed the two tasks in a quiet room, so that they could not be distracted. The IAT was created using Inquisit software (Draine, 1998), and the self-report questionnaire was created using Qualtrics.
29 3.5 Task 1
Task 1 was designed to measure the immediate, associative responses of the respondents to the different stimuli. Four categories were used in the IAT, presented in seven blocks. While the respondent was reading the two product descriptions, the researcher filled in the participants’ number and group number in the IAT. To overcome cognitive inertia, half of the respondents proceed through the block sequences 1-2-3-4-5-6-7 and the other half 1-5-6-7-2-3-4. If the respondent received group number one, the respondent proceed through the first block (1-2-3-4-5-6-7), and if the respondent received group number two, the respondents proceed through the other sequence. In figure 5, the difference in start is shown. In group number one, Pure Tea was linked with positive stimuli, and in group number two, Pure Tea was linked with negative stimuli.
In the first task, participants were told that different stimuli would be presented on the computer screen and that the respondents had to categorize each stimulus as belonging to the category shown in the upper left or upper right of the computer screen. This was possible by pressing the correct computer key, “E” for the left category and “I” for the right category. The four categories were Pure Tea and Orange Delight (target concepts) and positive and negative (target attributes). Participants were instructed to click quickly without being afraid of making a mistake. If a respondent categorized the stimuli incorrectly, a red “X” was shown to the respondent. The respondent could continue with the test by clicking on the correct computer key. The order of the trials within each block was randomized.
The stimuli words for Pure Tea and Orange Delight were measured during the pre-test. Respondents were asked to write down all associations they had with tea or orange juice. The most common words were chosen for the IAT. The stimuli for the target attributes, good and bad, were translate and retranslate directly from an existing IAT (Appendix B).
30
Figure 5: Start of the IAT. “Pulp” is the stimuli which respondents have to categorize.
The first block of the IAT was a practice stage, in which respondents were presented with different words that they had to link to either Pure Tea or Orange Delight. The second block was also a practice stage, in which respondents were presented with either positive or negative words that they had to link to either positive or negative. In the third block, the four categories were combined. Depending on the group number, Pure Tea was linked to positive or to negative. The fourth block repeated block three. In the fifth block, Pure Tea and Orange Delight were reversed on screen position; therefore, the fifth block was another practice stage, in which respondents had to link the different words to the new position of either Pure Tea or Orange Delight. In the sixth block, the two categories were again combined with positive or negative. Pure Tea and Orange Delight were reversed, whereas positive and negative remained on the same position. Therefore, each target concept was now combined with a different attribute. A list of the different stages is shown in Appendix D.
31 3.6 Task 2
Task 2 was designed to measure the respondents’ explicit attitudes toward Pure Tea and Orange Delight. Respondents were asked if they had a preference for one of the two, based on the earlier read product descriptions. Secondly, respondents had to rank six different statements on importance when purchasing a product. Three of these statements were transparency statements used from Visnawath & Kaufmann’s (2001) research. The other three statements were the opposite, thus non-transparency statements (see Appendix E).
32 4. Results
The previous chapter discussed the research design, pretest, and data collection of this study. In this chapter, the results of the data collection will be provided. First, the descriptive results will be discussed, then the implicit and explicit results will be presented.
4.1 Descriptive results
The total number of respondents who started the questionnaire was 91. However, the IAT excluded 6 respondents from the results due to latencies below 300 milliseconds or above 10,000 milliseconds (Greenwald et al., 2003). So the final sample consisted of 85 respondents. Respondents ranged in age from 18 to 63, averaging 34. Of all respondents, 39 were men (45.9%) and 46 were women (54.1%). The education level of the respondents was distributed as follows: 12 respondents (14.1%) finished or were still studying in high school, another 12 respondents finished or were still studying MBO, 33 respondents (38.8%) finished or were still studying HBO, and a total of 28 respondents (32.9%) finished or were still studying at the university.
The three conditions were filled by 85 respondents. In each condition, half of the respondents received group number 1 and the other half received group number 2. As explained earlier, this was done to overcome cognitive inertia. In the first condition, 29 participants, 14 males and 15 females, filled in the questionnaire. In the second condition, 28 participants, 10 males and 18 females, filled in the questionnaire. Finally, the third condition, was filled in by 28 participants, 15 males and 13 females. Tables 1a, 1b, and 1c show a comparison among the three conditions. This is to determine if the results in each condition differ significantly because significant differences could possibly influence the final results.
33
Table 1a
Comparison in Group Number between the Three Conditions
Group 1 Group 2 Total
Condition 1 15 14 29
Condition 2 14 14 28
Condition 3 14 14 28
Table 1b
Comparison in Age and Gender between the Three Conditions
Table 1c
Comparison in Education Level between the Three Conditions
High school MBO HBO University
Condition 1 4 4 13 8
Condition 2 2 3 12 11
Condition 3 6 5 8 9
All respondents received either group number 1 or 2. In each condition, the two numbers were counterbalanced to overcome cognitive inertia, which distorts individual IAT scores (Greenwald et al., 2003). To determine if there was indeed a difference between the two groups, an independent sample t-test was carried out. The test showed that there was a significant difference in D-scores between the mean in group one (M = 0.07, SD = 0.43) and the mean in group two (M = -0.3, SD = 0.45); t(83) = 3.94, p = 0.000). These results suggest that the order in which respondents completed the IAT influenced the outcome. The product that was first linked with positive stimuli was implicitly preferred more highly than the product that was first linked with negative stimuli.
4.2 Implicit results
All respondents completed the same IAT, only the group number (1 or 2) influenced the order in which they saw the IAT. Inquisit, the software program, aggregated and transformed the IAT data with the use of a D measure (Greenwald et al., 2003). The D
Male Female Age (mean)
Condition 1 14 15 39
Condition 2 10 18 31
34
measure is similar to Cohen’s d measure. However, Cohen’s d uses a pooled within-treatment standard deviation, whereas the IAT’s D measure uses a standard deviation, calculating the scores in both measurement blocks, independent of the condition of each score (Greenwald et al., 2003; Pantos & Perkins, 2012). Cohen’s d is calculated through the difference between means divided by the standard deviation (Cohen, 1988). The D-score takes more information into account than Cohen’s d and is, therefore, a better measurement (see table 2).
Table 2
IAT D-score Procedures
1 Delete trials greater than 10,000 milliseconds
2 Delete subjects for whom more than 10% of trials have latency less than 300 milliseconds
3 Compute the “inclusive” standard deviation for all trials in Stages 3 and 6 and likewise for all trials in Stages 4 and 7
4 Compute the mean latency for responses for each of Stages 3, 4, 6, and 7
5 Compute the two mean differences (MeanStage 6 – MeanStage 3) and (MeanStage 7 – MeanStage 4)
6 Divide each difference score by its associated “inclusive” standard deviation 7 D = the equal-weight average of the two resulting ratios
Note. From Greenwald et al. (2003, Table 4, p.214).
Data was eliminated from the IAT when there where latencies of less than 300ms (random responses) or more than 10,000 ms (lapses of concentration) (Greenwald et al., 2003). In this study, six participants were eliminated because they did not satisfy the requirements.
A D-score close to -1 indicated that the respondent had a strong preference for Orange Delight, whereas a score close 1 indicated that the respondent had a strong preference for Pure Tea. When the D score was between 0 and (-)0.15, little to no preference was measured, a D score between (-)0.15 and (-)0.35 indicated a small preference, a D score between (-)0.35 and (-)0.65 indicated a moderate preference, and a D score between (-)0.65 and 1 indicated a strong preference (Greenwald et al., 2003). The mean of the D scores of each condition is showed in table 3. To determine if there was a difference in D-scores among the three conditions, a one-way analysis of variance was conducted. The test showed F(2.82) = 3,056, p
35
= 0.05. This result shows that there was a significant difference in D-scores among the three conditions. A contrast coefficient test was conducted to determine if there was a significant difference between conditions one and two. It showed that there was a significant difference between conditions two and three (p = 0.01). No significant difference was found between conditions one and two (p = 0.25) or between one and three (p = 0.18).
Table 3
D-Score in Each Condition
Mean SD Indicates
Condition 1 -0.12 0.45 Little to no preference for Orange Delight Condition 2 -0.26 0.46 A slight preference for Orange Delight Condition 3 0.04 0.47 Little to no preference for Pure Tea
Figure 6: Graph of D-score
In condition 3, the control condition, the D score was near zero. This result was expected because both products were presented in the same way. The results for conditions one and two could indicate that participants had an automatic preference for orange juice, irrespective of the type of product (transparent or not transparent).
The distribution of the D-score is presented in figure 7. The grey bars show the participants that had an implicit preference for Orange Delight, and the black bars show the participants that had an implicit preference for Pure Tea. The white bars indicate no specific preference. The histogram was tested on skewness and kurtosis. Kurtosis measures the degree
36
to which scores cluster in the tails of frequency distributions (Field, 2009, p. 788), whereas skewness measures the symmetry of the frequency distribution (Field, 2009, p.794). It showed that the histogram has a skewness factor of 0.11 (Std. Error of skewness: 0.261), meaning that it is lightly right skewed. Values above 0.2 indicate great skewness (Hildebrand, 1986); therefore, this histogram is not extremely skewed. The kurtosis factor is –0.082 (Std. Error of kurtosis: 0.517), meaning that there is a platykurtic distribution, indicating that the histogram is flatter than a normal distribution with a wider peak.
Figure 7: Histogram of D-Scores
It was expected that consumers implicitly would have a preference for the non-transparent product because consumers try to avoid information that could lead to an undesired action or change of beliefs. The non-transparent product would, therefore, give consumers a positive feeling, resulting in an automatic preference. However, the results showed differently; therefore, hypothesis 1 is rejected.
37 4.3 Explicit results
Each respondent was asked which preference they had based on the earlier product descriptions. In condition 1, Orange Delight was presented as a transparent product and Pure Tea as a non-transparent product. In condition 2, it was the opposite. In condition 3, the control condition, both products were presented as non-transparent. The different stories used in the conditions can be found in Appendix C. In table 4, the respondents’ explicit preferences in each condition are shown. A score close to five indicates a strong preference for Orange Delight and a score close to one indicates a strong preference for Pure Tea.
Table 4
Explicit Preferences of the Respondents
Mean SD Indicates
Condition 1 3.97 1.26 Preference for Orange Delight Condition 2 2.14 1.16 Preference for Pure Tea Condition 3 2.79 1.10 No specific preference
These results show that in condition 1 and 2, the transparent product was more preferred than the non-transparent product. A one-way analysis of variance was conducted to determine if the means of the three conditions significantly differed. This result (F(2.82) = 16,346, p = 0.000) showed that there is indeed a difference in means between the three versions. A contrast coefficient test was conducted to determine if there was a difference between condition 1 and 2. It showed that there was indeed a significant difference between condition 1 and 2 (p= 0.000). Hypothesis 2 expected that consumers explicitly prefer the transparent product over the non-transparent product; therefore, hypothesis 2 is confirmed.
4.4 Purchase results
It was expected that purchase intentions are more strongly related to explicit attitudes than to implicit attitudes. This is because consumers could deliberately think about the different statements and form a conscious opinion. Therefore, it was expected that consumers
38
found transparency important in their purchase behavior. All respondents had to rank six different statements on importance when purchasing a product, where 1 was very important and 6 was least important. Three of these statements were transparency statements and the other three were the opposite. The transparency statements were seen as more important when purchasing a product and thus are more highly ranked than the non-transparency statements (see table 5). To determine if there was concordance in ranking among the respondents, a Kendall’s W test was executed. When Kendall’s W is close to 1, respondents totally agree in their ranking, whereas if Kendall’s W is close to 0, respondents totally disagree (Statistics
Solutions, 2013). Kendall’s W was shown to be 0.640 and was significant (p = 0.000),
indicating a large degree of agreement among respondents. Table 5
Descriptive Statistics of Valuation Statements
Mean ranking position (1 = highest, 6 = lowest) SD Information relevance 1.42 1.028 Quality mark 2.61 1.048
Open and honesty 2.92 0.862
Attractive look 4.05 1.112
Makes me happy 4.31 1.291
Celebrity promotion 5.69 0.756
Because the transparency statements were ranked more highly than the non-transparency statements, indicating the purchase intentions are indeed more strongly related to explicit attitude, hypothesis 3 is accepted.
39 4.5 Correlation between IAT results and explicit findings
To examine the relationship between the results from the IAT and the explicit findings, a correlation analysis was conducted. No significant correlation was found between the implicit and explicit results (r = -0.013, p = 0.90). In table 6, the results of the correlation analysis are presented.
Table 6
Means, Standard Deviations, and Correlations
Variables M SD 1
1 Explicit preference 2.96 1.38 - 2 Implicit preference -0.11 0.47 -.013
To determine if the demographic variables have an influence on implicit or explicit attitudes, a linear regression was used. It was found that the independent variable, gender, can significantly predict the outcome of the dependent variable, explicit preference: F(1.83) = 4.451, p = 0.038. Examining the adjusted R square shows 3.9% of the variance in explicit preference is explained by gender. Other descriptive variables, age or educational level, could not significantly predict the outcome of explicit attitudes. Implicit attitudes could not be predicted by any descriptive variables.
40 5. Conclusion
In the previous chapter, all data was analyzed and the different hypotheses were accepted or rejected. This chapter will first give a summary of the results, followed by theoretical and managerial contributions. Finally, a general conclusion will be presented, including limitations and suggestions for future research.
This study examined the influence of transparency on implicit and explicit attitudes and on purchase behavior. This study found that transparency had a positive influence on explicit attitudes, confirming hypothesis 2. In condition one and two, the transparent product was explicitly favored over the non-transparent product and in condition three, the control condition, no specific preference was noted.
It was expected that consumers implicitly preferred the non-transparent product over the transparent product. Because consumers try to seek positive information that supports their beliefs (Meissen et al., 1995), attractive, non-transparent information would positively influence a consumer’s attitude toward a product. In condition two, the non-transparent product, Orange Delight, was indeed implicitly preferred more highly than the transparent product, confirming hypothesis 1. However in condition one, the transparent product, Orange Delight, was implicitly preferred more highly as well, rejecting hypothesis 1. In condition three, the control condition, the implicit attitude was near zero, meaning that there was no specific preference. A significant difference was found between conditions two and three, indicating that hypothesis 1 is partially supported.
To measure purchase behavior, respondents had to rank six different statements on importance during the purchasing process. Results showed that respondents ranked transparency statements higher and, therefore, purchase intentions are indeed influenced by the outcome of explicit attitudes, confirming hypothesis 3.
41 5.1 Theoretical contribution and managerial implications
Only a few studies have addressed transparency in the context of marketing (Eggert & Helm, 2003). Therefore, this study contributes to scholarly work within this field. In the pre-test, two transparent and two non-transparent stories were tested on the different dimensions of transparency: access, comprehensiveness, relevance, quality, and reliability (Visnawath & Kaufmann, 2001). The transparent stories indeed scored higher on the five dimensions, confirming that these dimensions are transparency dimensions.
Several research studies already mentioned the importance of transparency (Bhaduri & Ha-Brookshire, 2011; Cohn & Wolfe 2012, 2013), only the influence on implicit and explicit attitudes was not yet researched. Other research already examined the difference between implicit and explicit attitudes (e.g., Devine, 1989; Greenwald & Banaji, 1995; Greenwald et al., 2003). However, within the topic of transparency, no research was yet conducted. Therefore, this study contributes to scholarly work about transparency and attitude.
Results showed that respondents explicitly preferred the transparent product over the non-transparent product. This result confirms that transparency is indeed an important factor in the decision making process and that consumers are more conscious about the environment (Bhaduri & Ha-Brookshire, 2011). With regard to hypothesis 1, consumers implicitly preferred a non-transparently marketed product over a transparently marketed product, a contradiction was found in the results. In condition one and two, Orange Delight was implicitly favored over Pure Tea. This might indicate that respondents preferred orange juice, irrespective of transparency. Prior experience with tea or orange juice may, therefore, have influenced the respondents’ implicit attitudes. Further research is necessary to confirm this. The different results between implicit and explicit attitudes show that they are two distinctive processes, confirming earlier the research of Devine (1989) and Greenwald and Banaji (1995). Messner and Vosgerau (2010) stated that implicit attitude might be a better predictor
42
of behavior than explicit attitude. This study showed differently, indicating that purchase behavior is influenced by the outcome of explicit attitude.
Half of the participants were placed into group number 1 and the other half into group number 2 to overcome the problem of cognitive inertia. In group number 1, Orange Delight was first matched with positive words and in group number 2, Pure Tea was first matched with positive words. Results showed that the product that was first linked with positive stimuli was implicitly more highly preferred than the product that was first linked with negative stimuli. This problem of internal validity was already mentioned in Greenwald et al., (2003) and thus basically means that the IAT may not be a reliable tool to measure implicit attitude and, therefore, contradicts previous research showing that the IAT is a valid measure (Maison et al., 2004).
The confirmation of hypothesis 3, the influence of explicit attitude on purchase behavior, confirms earlier research that mentioned the importance of good corporate behavior and credibility on purchase intentions (Creyer & Ross, 1996; Lafferty & Goldsmith, 1999). No correlation was found between implicit and explicit attitudes, contradicting earlier research, which did find a relationship between explicit measures and the IAT (Dovidio, et al., 1997, experiment 2; Wittenbrink, et al., 1997).
This study also offers implications from a managerial point of view. If a consumer explicitly has to make a choice between a transparent and a non-transparent product, consumers will choose the transparent product. Therefore, it should be recommended that companies be transparent. However, consumers do not necessarily search for transparent information because information might lead to unpleasant emotions, undesired actions, or changes of belief, which consumers try to avoid. Therefore, it is important for managers to know if being transparent would lead to those consequences for consumers. If transparency does lead to those consequences, companies should not be transparent. If being transparent
