1
University of Amsterdam
The influence of abstraction appeals and spatial distance appeals in
green advertisements on consumers’ purchase behavior
Master’s thesis
MSc. in Business Administration – Marketing track Academic year 2015-2016
Submission date: 22th of June 2016 Version: Final version
Author: Denise van Rijk (6070469) Supervisor: Anouar El Haji
2 STATEMENT OF ORIGINALITY
This document is written by Denise van Rijk who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 TABLE OF CONTENTS STATEMENT OF ORIGINALITY 2 ABSTRACT 5 1. INTRODUCTION 6 2. LITERATURE REVIEW 8 2.1 Construal Level Theory 8
2.1.1 Level of construal 8
2.1.2 Psychological distance 9
2.1.3 Conceptual research on CLT 10
2.1.4 Empirical research on CLT 12
2.2 Consumers’ environmental attitudes 14
2.2.1 Environmental involvement 14
2.2.2 Perceived cause importance 16
2.3 Ethical purchase behavior 17
2.4 Literature gaps & research question 17
2.5 Hypotheses 19 3. METHODOLOGY 23 3.1 Research design 23 3.2 Research sample 23 3.3 Method 24 3.4 Materials 24 3.4.1 Product characteristics 24 3.4.2 Green advertisements 25 3.4.3 Survey 28 3.5 Procedure 28
4
4. RESULTS 29
4.1 Descriptive statistics 29
4.2 Consumers’ purchase behavior 30
4.3 Explorative analysis 39
5. DISCUSSION 41
5.1 Summary of the findings 41
5.1.1 The influence of abstraction appeals and spatial distance appeals 41 5.1.2 The influence of consumers’ environmental attitudes 43
5.1.3 Additional influences 44
5.2 Theoretical implication 45
5.3 Managerial implications 47
5.4 Limitations and future research 47
5.5 Conclusion 49
REFERENCES 51
APPENDICES 54
Appendix A: Advertisements per condition 54
Appendix B: Pilot study: material 57
5 ABSTRACT
Green marketing is a relatively young field of study and marketing practitioners are always searching for the most effective appeals in advertisements. This study investigated the influence of abstraction appeals and spatial distance appeals in green advertisements on consumers’ purchase behavior. Abstraction and spatial distance appeals are derivatives of construal level and psychological distance; components of the Construal Level Theory. Abstraction appeal refers to the concrete vs. abstract way the text is written. Spatial distance appeal refers to the described physical distance between two events in the text: proximal vs.
distant. In this field-experiment a sustainable water bottle (the Dopper) was auctioned, and
1104 consumers participated in the auction. The text of four experimental advertisements was manipulated on abstraction level: concrete/abstract, and on spatial distance level: low/high. One control advertisement consisted of neutral appeals. Furthermore, the moderating effects of consumers’ environmental attitudes, in particular environmental involvement and perceived cause importance, were examined as well. The results showed that concrete appeals had a more positive influence on consumers’ purchase behavior than abstract appeals. The low and high spatial distance appeals did not influence consumers’ purchase behavior. Under certain conditions, environmental involvement and perceived cause importance positively moderated the relationship between the appeals and purchase behavior. For the further development of the CLT in a green context and recommendations for marketing practitioners, it is relevant to investigate the relationship further. In future research, the appeals must be manipulated more sophisticated and the experimental setting must be more neutral.
Key words: green marketing, Construal Level Theory (CLT), spatial distance, environmental involvement, perceived cause importance, purchase behavior, willingness to pay (WTP), sustainable products.
6 1. INTRODUCTION
In recent years people have developed a preference for the maintenance of a healthy and conscious lifestyle, which considers the environment and others as well (Ambec & Lanoie, 2008; Verbeek & Boelhouwer, 2010). As a consequence the concept of sustainability has been moved into the foreground of daily life, and companies are expected to respond to this. To sell their sustainable products companies use green marketing tactics (Agrawal & Das, 2013). Green marketing is a relatively young field of study and green consumer behavior has been barely explored (Yang et al., 2015). Insight in effective green marketing tactics is valuable, because green marketing will benefit companies in the long run (Agrawal & Das, 2013).
In the literature no clarity exists about the most effective appeals in green marketing. Many studies examined the influence of abstraction appeals and spatial distance appeals (Darley & Smith, 1993; Lii et al., 2013; Goldsmith et al., 2014; White et al., 2011; Yang et al., 2015). These appeals are derivatives of construal level and psychological distance; components of the Construal Level Theory (CLT). Abstraction appeal refers to the concrete vs. abstract way the advertisement text is written. Spatial distance appeal refers to the described physical distance between two events in the advertisement: proximal vs. distant. Some researchers recommend to design advertisements with concrete appeal (Darley & Smith, 1993; Lii et al., 2013), but others suggest the opposite: abstract appeals should influence consumer behavior more easily (Goldsmith et al., 2014; White et al., 2011; Yang et al., 2015).
Since no clarity exists in the literature about the most influential appeals, this study aims to bridge this gap. In particular, the combination of abstraction appeals and spatial distance appeals in a green context is barely examined before. This combination is especially valuable in a green context, since companies’ sustainable activities find place on both physically close and distant locations. The aim of this study is to answer the research
7 question: What is the influence of abstraction appeals and spatial distance appeals in green advertisements on consumers’ purchase behavior?
This is investigated in a field-experiment with a sample of 1104 Dutch-speaking consumers. Via an online Vickrey auction consumers’ willingness to pay is measured for a sustainable water bottle: the Dopper. In the auction consumers are exposed to one of the five green advertisements. The advertisements are manipulated on abstraction level and spatial distance level (or consist of neutral appeals) to determine the influence of the appeals on consumers’ purchase behavior. In addition, the moderating effects of consumers’ environmental attitudes, in specific environmental involvement and perceived cause importance, are examined as well.
When findings show more positive effects of one appeal over the other or for a combination of two appeals, both theory and practice will benefit from the findings. The CLT can be developed further in a green marketing context, and practitioners can use the findings to improve their marketing and communication materials.
The remainder of this article is structured as follows. In the next section an overview of the current literature is presented and the hypotheses are developed. In the third section the methodology of our field-experiment is outlined. The results of the auction data are presented in the fourth section. In the last section a summary of the findings, the theoretical- and managerial implications and the limitations are discussed, followed by the final conclusion.
8
2. LITERATURE REVIEW
This section presents an overview of the current insights in the main concepts of this study. First, the Construal Level Theory (CLT) and its conceptual research are described, followed by a review of the CLT’s experimental research.Second, research about consumers’ environmental attitudes in the effectiveness of green advertisements is discussed, and ethical purchase behavior is reviewed. Afterwards, the discovered gaps in the literature are mentioned, followed by the research question and the defined hypotheses for this study.
2.1 Construal Level Theory
The consumer psychology discipline aims to understand an individual’s decision making and evaluation process, in order to identify the most influential factors in consumers’ choices (Trope et al., 2007). Construal Level Theory (CLT) explains consumer evaluation and consumer behavior via the linkage between construal level and psychological distance (Trope et al., 2007). The CLT’s central premise is that an individual’s representation of information (construal) depends on the degree of psychological distance: proximal vs. distant (Leiser et al., 2008). More specific the theory states that “individuals use concrete low-level construals to represent near events, and abstract high-level construals to represent distant events” (Trope et al., 2007, p. 83). In the next paragraphs, the concepts construal level and psychological distance are explained.
2.1.1 Level of construal
Construal refers to the subjective way how individuals perceive, interpret and
understand the world around them, i.e. one’s representation of information (Ross, 1987).
Concrete low-level construals are explained as contextual, relatively unstructured
representations with incidental and inferior features of an event. Abstract high-level construals are explained as more general, relatively structured representations with more
9 superordinate core features of an event (Trope et al., 2007). Thus, high-level construals are more structured and simpler, but less rich and less detailed than low-level construals (Leiser et al., 2008). For example, in the event “studying at university” a low-level construal is “reading scholarly books”, since this is concrete behavior and inferior to the event “studying at university”. At the same time, a high-level construal is “investing in the future”, because this describes abstract, more general (less detailed) behavior and is superior to the event. Another example is “going to the cinema”, whereby “booking tickets online” presents a low-level construal and “being entertained” a high-low-level construal. So, the difference lies in the approached construal level when a representation of the event is made in an individual’s mind. In brief, low-level construals are concrete, complex, contextualized and subordinate representations focusing on ‘how’, whereas high-level construals are abstract, simple, decontextualized and superordinate representations focusing on ‘why’.
2.1.2 Psychological distance
Psychological distance is defined as: “the subjective experience that something is
close or distant from the self, here, and now” (Trope & Liberman, 2010, p. 440). The CLT presumes that psychological distance plays an important role in consumer evaluation and behavior; psychologically distant events are processed in a more abstract way (i.e. high-level construal), and conversely psychologically close events are processed in a more concrete way (i.e. low-level construal) (Trope et al., 2007). Psychological distance can be applied on multiple levels and these forms are described shortly. (1) Spatial distance: the actual geographical distance between the event and the reference point (Lii et al., 2013). According to the CLT, an individual uses a more concrete mindset for events located proximally (“at your home”; “in your country”) and a more abstract mindset for events located remotely (“500 miles from here”; “in another country”). (2) Temporal distance: the actual distance in time between the event and the reference point (e.g. hours, days, years) (Lii et al., 2013). An
10 individual should use a more concrete mindset for events nearby (“tomorrow”; “next week”) and a more abstract mindset for distant events (“next year”; “in ten years”). (3) Social
distance: the level of similarity between the observer and the endorsed person/firm of the
event (Lii et al., 2013). The CLT presumes that an individual uses a more concrete mindset for persons more similar to oneself (“colleague”; “fellow citizen”) and a more abstract mindset for persons less similar to oneself (“criminal”). (4) Hypothetical distance: the event’s likelihood to take place (Lii et al., 2013). An improbable event is seen as more distant than a probable event; when the probability of the event decreases, the psychological distance increases (Trope et al., 2007). The theory states that an individual uses a more concrete mindset for events that are likely to occur (“voting by elections in the Netherlands”) and a more abstract mindset for events that are not likely to occur (“being attacked by a lion in the Netherlands”).
2.1.3 Conceptual research on CLT
So far, it has been assumed that the construal (abstraction) level and psychological distance level are related to each other. Trope and Liberman (2007; 2010) have been frontrunners in the conceptual development of the CLT, and they have been aimed to gather deeper insights in this relationship. First, the scholars emphasize that construal levels and psychological distance are associated, but the concepts are definitely not the same: “Psychological distance refers to the perception of when an event occurs, where it occurs, to
whom it occurs, and whether it occurs. Construal level refers to the perception of what will
occur: the processes that give rise to the representation of the event itself” (Trope & Liberman, 2010, p. 442). About the nature of the relationship, Trope and Liberman (2007) propose a bi-directional association between construal level and psychological distance. Bi-directionality stands for functioning in both directions: construal level influences psychological distance as well as psychological distance influences construal level, see
11 Figure 1. Thus, when the abstraction level increases (decreases), the perception of psychological distance increases (decreases) and vice versa: as psychological distance increases (decreases), the construal level becomes more abstract (concrete) (Trope et al., 2007). Consequently, Trope et al. (2007) recommend to create a ‘fit’ between the construal level and the level of psychological distance in organized representations, because they found that fitting representations affect individuals’ responses most easily.
Second, Trope and Liberman (2010) highlight the strength of the relationship among proximal and distant events. In first instance, they state that associations among proximal events could have the same strength as associations among distant events. Nonetheless, it is reasonable that proximal events have more in common than distant events, since psychological distance is egocentric and initiated from the self, here and now (Trope & Liberman, 2010). Since Trope and Liberman (2010) argue for a greater commonality among proximal events in comparison with distant events, this supports a stronger interrelationship between the former than the latter.
Third, Trope and Liberman (2010) noticed that the different psychological distances can be interrelated. The scholars gave an example of a situation wherein spatial distance and social distance are interrelated: one chooses a spatially more distant seat from another to show social distancing from that person. After the theoretical conceptualization of CLT, it is relevant to systematically analyze the current empirical research on CLT.
12 2.1.4 Empirical research on CLT
Many researchers have studied the CLT in an empirical setting and multiple studies have supported the relationship between construal levels and psychological distance (Trope et al., 2007). Since this study focusses on construal level in combination with spatial distance, it is relevant to inspect prior empirical research towards these concepts. Fujita et al. (2006) find a relationship between construal and spatial distance in a social context. In two experiments, Fujita et al. (2006) tested whether an increase in spatial distance of social events would increase the abstraction level of an individual’s construals. In their first study the social event is helping a friend to move to a close versus distant location, and in the second study the social event is watching a video of a conversation by students from a closely versus distantly located university. The results showed that an increase in spatial distance leads to a more abstract, global mindset: a high-level construal (Fujita et al., 2006). In first instance, it is valuable to our study to find support for the relationship between construal levels and spatial distance, although the study of Fujita et al. (2006) focused on social events rather than on marketing and consumer behavior.
Nevertheless, some researchers studied the CLT in an (green) advertising context, since construal levels in particular have often been applied in messages via abstract and concrete appeal. In advertisements, concrete appeal is defined as rich, detailed information with concrete, specific product feature descriptions (i.e. low-level construals) and abstract
appeal consists of ambiguous wording and abstract, unspecific product feature descriptions
(i.e. high level-construals) (Yang et al., 2015). In prior advertising research, the primary held thought was that concrete appeal in a message had more influence on consumers’ responses than abstract appeal (Darley & Smith, 1993). This is in line with current findings from Lii et al. (2013), who studied the effects of the sustainable marketing of multiple CSR initiatives on consumer evaluations with spatial distance as moderator. The scholars stated that concrete appeals would have a more positive effect of consumers’ evaluations in comparison to
13 abstract appeals in a CSR context. The findings showed that spatial distance served as a moderator in the relationship between CSR initiatives and consumers’ evaluations (attitude towards brand and perceived credibility of campaign). When a consumer perceived low spatial distance by the environmental cause or brand, the effect of the CSR activity on the consumer’s evaluation was higher (Lii et al., 2013). Thus, a message seems to be most effective when represented in a concrete, low spatial distance perspective.
Though, some researchers suggested otherwise: under some conditions, abstract appeals seemed to have a more positive influence on consumers’ responses than concrete appeals (Goldsmith et al., 2014; White et al., 2011; Yang et al., 2015). For example, Yang et al. (2015) studied how consumers were encouraged in purchasing green products by the manipulation of abstraction appeal (concrete vs. abstract) in three experiments. They find that abstract (vs. concrete) appeals were more effective in evoking consumers’ purchase intentions for green products. However, this relationship was moderated by the activation of ‘benefit of other’ and ‘collective level of self’. Thus, this finding weakens the support for a direct causal relationship between abstract appeal and purchase intention. In addition, Goldsmith et al. (2014) studied the effects of construal level (abstraction appeals) on the purchase intention of green products in three experiments. Their findings show that consumers with an abstract mindset were more likely to choose for a sustainable product and were willing to pay more for it, in comparison with the concrete mindset consumers. A moderator in this relationship was ‘self-transcendent benefit’ (Goldsmith et al., 2014). Unfortunately, in the studies of Yang et al. (2015) and Goldsmith et al. (2014) psychological distance was not taken into account in the research design.
Overall, in the current literature no consistency has been found for the determination of the most effective appeals. For the design of a green message, contradicting recommendations are given regarding abstraction appeal. Research towards the effectiveness of abstraction appeals in combination with spatial distance appeals, particularly in a green
14 advertising context, has not been executed often. Beyond that, the small number of experimental studies investigated the influence of the CLT’s components mostly as moderator instead of independent variable. A need arises to investigate the direct effects of the CLT’s components. Thus, it is important that the influences of abstraction appeals and spatial distance appeals can be determined separately. In this study both appeals are manipulated in order to test their direct effects. Though, it is reasonable that influences outside the message – such as consumers’ attitudes – play a role in the effectiveness of a green message as well. Since this study focuses on green marketing, consumers’ environmental attitudes are taken into account in the research design. In the next paragraph consumers’ environmental attitudes are reviewed, in particular environmental involvement and perceived cause importance.
2.2 Consumers’ environmental attitudes
As mentioned before this study focuses on the influences of construal level and spatial distance on consumers’ purchase behavior. Next to the influence of marketing tactics, individual characteristics can influence purchase behavior as well (Hinnen et al., 2015). In order to get a more realistic view of consumer behavior, this study examines the influences of some individual characteristics as well, in particular consumers’ environmental attitudes. Because of this the effectiveness of green advertisements can be determined more accurately. The next paragraphs analyze two forms of environmental attitudes: environmental involvement and perceived cause importance.
2.2.1 Environmental involvement
It is reasonable that an individual’s environmental involvement can positively influence the impact of a green message on consumers’ purchase behavior. Explained
15 otherwise, the more involved a consumer is with the environment, the more likely he/she is to transfer this attitude into real consumption behavior. Nevertheless, in the current literature discrepancies exist regarding the influence of environmental involvement on consumer behavior, since several studies showed different effects. To start with, Haytko and Matulich (2008) studied the effects of green advertising on the attitudes and environmental behaviors of different consumer groups, segmented on their level of environmental involvement. The findings showed that green advertising had more positive effects on high-involved consumers’ attitudes and behavior than on low-involved consumers (Haytko & Matulich, 2008). D’Souza and Taghian (2005) found the same results, when investigating the differences in green advertisement attitudes between low-involved and high-involved consumers. These findings are in line with the natural reasoning that individuals respond more easily when they feel (environmentally) committed to a case. However, other studies contradict this assumption and showed that green advertisements had more impact the other way around. For example, Schuhwerk and Lefkoff-Hagius (1995) examined consumers’ responses to different green advertisements for a green laundry detergent. They found that
low-involved consumers responded more positively towards the green advertisement than high-involved consumers.
In short, in the literature two-sided effects of environmental involvement are found. Still, it is relevant to understand how different segments, like differently involved consumers, can be influenced in their behavior. Perhaps environmental involvement is a too broad concept for direct segmentation, since it is arguable that one is involved differently with distinctive environmental causes. Following this line of thought, it is interesting to study one’s perceived cause importance (PCI) as well. PCI refers to one’s attitude about a particular environmental cause, instead of their general environmental attitude. In order to get a better understanding of consumers’ environmental attitudes, this study reviews the concept of PCI.
16 2.2.2 Perceived cause importance
In the green marketing context of this study, perceived cause importance (PCI) refers to the importance of the underlying environmental- or societal issue itself, in the eye of the consumer (Kronrod et al., 2012). Note this does not contain one’s opinion about the desired behavior or whether this behavior leads to positive consequences for the environment or society (Kronrod et al., 2012). The PCI is for example, one’s considered degree of importance in the fact that a large amount of people are homeless. Hence, it is not about the importance to donate money to these people or charity to try to diminish the amount of homeless people. It contains one’s perception of the importance of the cause and nothing more. Previous studies show that the greater extent to which one perceives the cause to be important, the more one listens or follows a promoting message of that cause (Clark, 1998; Cleveland et al., 2005). PCI can be seen as an influential factor on consumers’ green (consumption) behavior, and therefore this study takes this variable into account.
Nevertheless, consumers’ environmental attitudes alone, in particular environmental involvement and PCI, cannot simply declare the whole story. Other attitudes and demographics can also influence consumer’s purchase behavior, such as familiarity (with the product), liking (of the product), age and gender.
This section reviewed the main concepts that potentially influence purchase behavior in this study. Eventually, this study aims to get a better understanding of consumers’ purchase behavior, in particular ethical purchase behavior. Consumers’ motivations differ by the purchase of sustainable goods compared to regular goods. Therefore, it is relevant to understand the latest research and trends about ethical purchase behavior as well. In the next paragraph a short review of this concept is given.
17 2.3 Ethical purchase behavior
Consumers can express their environmental and societal concerns via their purchase behavior. For example, consumers can create demand through the purchase of sustainable goods or diminish demand for non-sustainable goods via boycotting (De Pelsmacker et al., 2005). Ethical purchase behavior consists of the procurement of a product or service that concerns a specific ethical issue (e.g. environment, human rights, animal wellbeing, etc.) that the consumer has freely chosen (Doane, 2001). Ethical consumption can benefit both the natural environment (eco-friendly products) and people (products made without child labor). It is stated that the market of sustainable products is growing, what logically follows an increase in demand (De Pelsmacker et al., 2005). The situation turned out to be more nuanced; yet the direction of consumers’ demand is going both ways. Furthermore, more people claim they buy sustainable products, but the real market shares remain pretty low (De Pelsmacker et al., 2005). Explanations for this situation are dispensed, what causes a need for more research towards ethical purchase behavior. This study can contribute to a better understanding of the efficiency of green marketing at the purchase of sustainable products.
2.4 Literature gaps & research question
Although a start has been made in the research about the effectiveness of green advertisements on consumer behavior, the demand for more extensive insights is still growing (Leonidou et al., 2011). In particular, the following two issues need to be addressed in the near future. First, for the success of green marketing a proper execution of the advertising appeals is crucial (Yang et al., 2015). Thus, there is a need for clarity about the effective appeals in green messages. The research is narrow upon this, since the influence of abstraction appeals on green (purchase) behavior has been barely explored (Yang et al., 2015). It is especially valuable to investigate the direct influences of the appeals in the message, instead of testing the moderating influences again.In this way, support for the CLT
18 can be strengthened or weakened and possibly a further development of the theory could take place. Subsequently, it is interesting to examine the influence of abstraction appeals in
combination with spatial distance appeals. In a green context, the concept of spatial distance
is crucial since because CSR-activities take place everywhere; spatially close and distant. The current literature lacks in the understanding of which spatial distance level influences consumers’ purchase behavior more positively. In addition, when studying purchase behavior the influences of environmental involvement and PCI need to be taken into account as well, since prior research emphasized the effects of both. The question arises whether (and when) consumers rely on their environmental attitudes and/or on the CLT appeals, when confronted with green advertisements in a purchase situation.This study aims to provide more clarity to the effectiveness of abstraction appeals and spatial distance appeals in green advertisements. Second, the methods used in prior research by investigating the effects of green advertisements were often vulnerable with regard to their external validity. Many studies operationalized consumers’ purchase behavior through the self-report of attitudes and purchase intentions (survey questions) rather than examining real-life purchase behavior. In first instance, consistent evidence is found for the positive influence of attitudes on behavior (De Pelsmacker, 2005). However, attitudes alone are weak predictors for real behavior which is known as the attitude-behavioral gap (Vermeir & Verbeke, 2006). Thus, research consisting of survey studies alone lacks external validity. Field-experiments can help through their increased realism in comparison with survey studies. Furthermore, field-experiments can test causal relationships since the variable manipulation takes place in a real-life setting. Particularly for green advertisements, it is valuable to study the effects in real life situations since intuitively all people are concerned about the environment. In surveys most people express they do care because of social desirability bias: the tendency to give social desirable answers. However, real life situations show that fewer people actually translated their self-reported attitudes to the actual purchase of sustainable products (Yang et al., 2015). In order
19 to eliminate this bias, it would be valuable to investigate the effects of green advertisements in a real life consumer buying situation. This study can make a contribution to that point.
This research purposes to address the two issues mentioned above by studying the effectiveness of green advertisements with appeals based on the CLT components. In a real consumer purchase situation, the effects of green advertisements with abstraction appeal and spatial distance appeal are investigated on purchase behavior. Therefore, the research question reads as follows: What is the influence of abstraction appeals and spatial distance appeals in green advertisements on consumers’ purchase behavior?
2.5 Hypotheses
After a review of the literature and the formulation of the research question, in this paragraph the hypotheses are developed. To start with the CLT, previous studies show that the manipulation of construal level (abstraction appeals) and spatial distance (appeals) in advertisements influenced consumers’ responses. Our first expectation is that green advertisements with abstraction appeals and spatial distance appeals influence consumers’ purchase behavior more positively than green advertisements with neutral appeals. The first hypothesis is stated as:
H1: Consumers exposed to abstraction appeals and spatial distance appeals in the green advertisement are willing to pay more than consumers exposed to neutral appeals.
Second, the CLT states that individuals use a more concrete mindset for close events, and a more abstract mindset for distant events (Trope et al., 2007). Trope and Liberman (2007; 2010) recommend to create a correct fit (in appeal level) in a message to achieve the most positive effects. A correct fit is assumed when both appeals are manipulated on the same level: both close (concrete) or both distant (abstract) (Trope et al., 2007; Trope &
20 Liberman, 2010). In this study, a correct fit in appeal level is applied to the advertisements with concrete, low spatial distance appeal and abstract, high spatial distance appeal, since here the appeals act on the same level. An incorrect fit in appeal level is applied to the advertisements with concrete, high spatial distance appeal and abstract, low spatial distance appeal, since these appeals act on a different level. Therefore, we expect:
H2: Consumers exposed to a correct fit in appeal level (in the advertisement) are willing to pay more than consumers exposed to an incorrect fit in appeal level.
For the development of the third hypothesis, a distinction is made in the effects of advertisements with a correct fit in appeal level and advertisements with an incorrect fit in appeal level. In case of correct fit, Trope and Liberman (2010) state that people care more about proximal events (“self, here, now”) than about distant events. In addition, since this study’s auction requests consumers to bring out a bid directly, it is reasonable that one activates a more concrete mindset in such a situation. These two assertions give more weight to concrete appeal over abstract appeal. Thus, for the correct fit advertisements is expected:
H3a: Consumers exposed to concrete, low spatial distance appeal in the advertisement are willing to pay more than consumers exposed to abstract, high spatial distance appeal.
In case of advertisements with an incorrect fit in appeal level it might be more nuanced. On the one hand, we expect a more favorable effect of concrete appeal over abstract appeal, since concrete is more proximal. But on the other hand, this rule can be applied to spatial distance as well: low distance is expected to be more influential than high distance. Both statements rely on the same assumption that people care more about proximal events than about distant events (Trope & Liberman, 2010). This reasoning from both sides gives no
21 preference in direction, and therefore we expect no differences on consumers’ purchase behavior among the incorrect fit advertisements:
H3b: Consumers exposed to concrete, high spatial distance appeal do not significantly differ in their willingness to pay from consumers exposed to abstract, low spatial distance appeal.
In addition, it turned out to be valuable to investigate some individual characteristics as well. It is possible that consumers’ environmental attitudes can moderate the relationship between abstraction and spatial distance appeals on consumers’ purchase behavior. In the current literature section particularly environmental involvement and PCI popped out as potential influencers in the effectiveness of green advertisements. Despite the contradicting findings in the studies towards the effects of environmental involvement, we follow the reasoning of Haytko and Matulich (2008) and D’Souza and Taghian (2005). These scholars state that people respond more easily when they feel involved to a case. Thus, we expect that the appeals in combination with a relative higher environmental involvement influence consumers’ purchase behavior more positively:
H4a: Environmental involvement has a positive moderating influence on the relationship between abstraction and spatial distance appeals (in the advertisement) and consumers’ willingness to pay.
For PCI also a positive moderating effect is expected, since we rely on the statement that the higher an individual’s perceived cause importance, the more one listens or follows a promoting message of that cause (Clark, 1988; Cleveland et al., 2005). Thus, we expect that the appeals in combination with a relative higher PCI influence consumers’ purchase behavior more positively:
22
H4b: Perceived cause importance has a positive moderating influence on the relationship between abstraction and spatial distance appeals (in the advertisement) and consumers’ willingness to pay.
Note that in the analysis of this study the subsequent influences of familiarity (with the product), liking (of the product) and some demographics are examined as well, but no hypotheses are stated for these concepts.
23 3. METHODOLOGY
After the literature review that discussed the main concept of our study, this section provides the methodology. First, the research design and the research sample are outlined, and afterwards a description of the method, the materials, the manipulation check and the procedure of the experiment are given.
3.1 Research design
This study conducted a field-experiment consisting of four experimental conditions and one control group. In five different green advertisements (conditions), appeals were manipulated on abstraction level (concrete vs. abstract) and spatial distance level (low vs. high). Participants were randomly assigned to one of the following conditions: Condition 1 [Concrete & low spatial distance]; Condition 2 [Concrete & high spatial distance]; Condition 3 [Abstract & low spatial distance]; Condition 4 [Abstract & high spatial distance]; and Condition 5 [Neutral]. The influence of the different appeals was examined on consumers’ willingness to pay-level (i.e. consumers’ bids) for a sustainable product in an online auction of Veylinx.
3.2 Research sample
About 3200 Dutch-speaking panel members of Veylinx were invited per email to participate in the auction. The response rate was 34.5 percent, since 1104 consumers actually made an offer in the auction designed for this study. The sample (N = 1104) consisted of males (49.8 percent) and females (50.2 percent), with an average age of 42.5 years old. Participants’ highest followed educational level consisted of: no education for 0.2 percent;
primary education for 1.1 percent; secondary education for 7.4 percent; intermediate vocational education (MBO) for 24.1 percent; applied sciences (HBO) for 32.4 percent; and university for 34.7 percent. In March 2016, 16.8 percent of the participants was a student.
24 3.3 Method
This field-experiment examined consumer purchase behavior via an online auction at Veylinx. Veylinx is a research platform developed to study how people value products and services, by scrutinizing consumers’ willingness to pay in online auctions. On a weekly basis, panel members of Veylinx receive an invitation (via e-mail) to bring out their bid for a particular product or service anonymously. The products and services are actually sold to the participants of the auction, and because of this we have access to real (consumer) data. Veylinx uses the Vickrey auction method for their online auctions. This method consists of a sealed-bid (blind) auction, where participants bid simultaneously for the same product or service (Lucking-Reiley, 2000). In Vickrey auctions the winner (i.e. the highest bidder) buys the product or service for the price of the highest losing bid; thus a price always lower than their own bid (Lucking-Reiley, 2000). This method is especially valuable since consumers are rewarded to not bid strategically, but to conform to their true value of the product or service. The Veylinx consumer panel consists of more than 3200 subscribers. Subscribers know Veylinx as a research platform, but they are not aware of the different conditions when participating in an auction. This guarantees a blind method. Furthermore, Veylinx randomly assigns participants to the different conditions. The outcome data of Veylinx’ auctions is consumers’ willingness to pay for the product or service, i.e. consumers’ bid amount in euros. The higher the bid is, the more one values the product in terms of money.
3.4 Materials
3.4.1 Product characteristics
The product chosen to auction was the Dopper: a sustainable, design water bottle for drinking water. Dopper’s mission is communicated as: “striving for a world where people are aware of the environment, where we actively reduce the amount of single-use plastic, and where everybody – close or distant from our homes – has access to safe drinking water”
25 (www.dopper.com). To achieve this mission Dopper focuses on three pillars: (1) “make people aware of the impact of single-use plastic waste”; (2) “develop sustainable design bottles for drinking water”; (3) “support drinking water and sanitation projects worldwide” (www.dopper.com). The recommended retail price is between 12.50 and 17.50 euros per bottle. In the Netherlands the Dopper became quite popular, due to the simple and concrete way one can support the environmental wellbeing and the easiness to expose where one stands for (Algera, 2013). We had chosen to study the Dopper because this product involves both close and distant spatial distance in the company’s CSR-activities. Therefore, it is possible to manipulate spatial distance in the green advertisements on both levels. In case of Dopper, close spatial distance consists of the reduction of single-use of plastic water bottles in the consumer’s direct environment. Remote spatial distance consists of the support of water and sanitation projects in Nepal. For this study ten identical “cool blue” Doppers were auctioned. The ten highest bidders (winners) all paid the price of the highest losing bid, thus the eleventh highest bid in this auction.
3.4.2 Green advertisements
The text of the green advertisements was manipulated on levels of abstraction appeal and spatial distance appeal. All other features of the advertisement were completely equal among the conditions, see Appendix A. In the advertisement texts concrete appeal is presented pragmatically focusing on ‘how’, and abstract appeal is presented idealistically focusing on ‘why’. In case of low spatial distance appeals is referred to one’s direct environment, and by high spatial distance appeals is referred to Nepal, which has a distance of 7000 kilometers from the Netherlands. The neutral message was designed without any form of spatial distance and with a moderate level of abstraction appeal. The used texts per condition are exhibited in Figure 2, and an example of a green advertisement is given in Figure 3.
26 Figure 2. Texts used in the different green advertisements.
Condition 1: Concrete & low spatial distance
“A more sustainable world starts with yourself! By using the sustainable water bottle Dopper you reduce single-use plastic waste in your direct environment.”
Condition 2: Concrete & high spatial distance
“A more sustainable world starts with yourself! By the purchase of the sustainable water bottle Dopper you support projects for safe drinking water and sanitation in Nepal.”
Condition 3: Abstract & low spatial distance
“A step towards a more sustainable world! It is beneficial for the environment to reduce the amount of waste. Dopper strives to a reduction of single-use plastic waste in the direct environment.” Condition 4:
Abstract & high spatial distance
“A step towards a more sustainable world! It is important to increase the access to safe drinking water for people all over the world. Dopper strives to improved drinking water- and sanitation services in Nepal.” Condition 5:
Neutral
“Dopper labors for a more sustainable world! Dopper wants to make people aware of the impact of single-use plastic waste and want to inspire people to create change by themselves.”
27 Figure 3. Advertisement of Condition 1 [Concrete & low spatial distance].
Manipulation check. A pilot study was executed to check whether the manipulation of
the appeals in the advertisements was successful. The levels of abstraction appeal and spatial distance appeal were evaluated in a small sample (N = 18). The questionnaire used is appended, see Appendix B. On a five-point Likert scale respondents were asked to evaluate the five texts on the level of abstraction appeal (1 = concrete, 5 = abstract) and the level of spatial distance appeal (1 = close, 5 = distant). The answers given were analyzed through one-sample t-tests: per level and per condition, and a significance level of p < .05 was maintained for all tests. The results of the pilot study are appended, see Appendix C. The results show that most of the abstraction appeal levels (low/high/neutral) and spatial distance appeal levels (low/high/neutral) were evaluated correctly. Therefore, we considered these appeals as manipulated sufficiently and no adaptation was needed. However, the results of the evaluations of spatial distance level in Condition 2 and abstraction level in Condition 3 seemed to be more nuanced. In Condition 2, the manipulated high level of spatial distance
28 turned out to be marginally significant only at level p < .10. Beyond that, the manipulated high level of spatial distance in Condition 2 was exactly equal to the high level of spatial distance in Condition 4, since both appeals refer to water projects in Nepal. Note that in Condition 4 this spatial distance manipulation was found to be significant. These results made us decide to not change the spatial distance appeal of Condition 2. In Condition 3, the manipulated high level of abstraction appeal turned out to be not significant at p < .05. Since the p-value found was located closely to the significance level boundary (p = .052), we decided to change neither the abstraction appeal in Condition 3.
3.4.3 Survey
Participants who confirmed their bid were also required to fill out the survey. The survey consisted of the following statements: (1) I feel involved by the wellbeing of the
environment; (2) The described environmental cause is important to me; (3) I was familiar with the Dopper before; (4) I evaluate the Dopper as follows. Question 1, 2 and 4 needed to
be answered on a five-point Likert scale (1 = totally positive/agree, 5 = totally negative/ disagree) and question 3 required a yes/no answer.
3.5 Procedure
On the 29th of March 2016, an e-mail invitation was sent to the consumer panel of Veylinx with the request to participate in an online auction. When one activated the invitation automatically one of the green advertisements of the Dopper was showed during a time slot of five minutes. During this time participants could bring out a bid starting from 0 euros till unlimited. After the bid confirmation, participants needed to fill out the survey: the last part of the experiment. The next day we closed the auction and participants received an e-mail with the auction’s outcome information: the highest bidder bet 20.50 euros, but all the ten winners paid just 12.50 euros (the price of the eleventh highest bid, i.e. the first losing bid).
29
4. RESULTS
This section describes the analysis and results of this study. To start with, the descriptive statistics of the hypothesis variables are presented and afterwards the analysis to test the hypotheses are described. In addition some explorative analyses took place. Note that we computed a new variable that contained the logarithm of ‘bid amount’, and this variable was used as outcome variable in all conducted tests.
4.1 Descriptive statistics
The descriptive statistics for consumers’ willingness to pay (WTP)-level per condition are presented in Table 2. Table 3 and Table 4 show the descriptive data for consumers WTP-level per dimension of Environmental Involvement and Perceived Cause Importance. The sample sizes (N), means and standard deviations (SD) are presented, and dependent on the variable the minimum and maximum values or the percentages.
Table 2. Descriptive statistics of WTP-level (in euros) per condition.
Condition N Mean SD Min. Max.
C1: Concrete, low spatial distance 251 4.51 2.56 0 8 C2: Concrete, high spatial distance 210 4.30 2.72 0 7 C3: Abstract, low spatial distance 200 3.94 2.71 0 7 C4: Abstract, high spatial distance 229 4.01 2.83 0 7
C5: Neutral 214 4.30 2.78 0 7
Total 1104 4.22 2.72 0 8
Table 3. Descriptive statistics of WTP-level (in euros) per Environmental Involvement dimension.
Environmental Involvement
N Mean SD Percent Valid Percent
Cumulative Percent Valid Totally agree (1) 377 4.36 2.71 34.1 34.2 34.2
Agree (2) 389 4.46 2.61 35.2 35.3 69.5 Neutral (3) 228 3.89 2.78 20.7 20.7 90.2 Disagree (4) 66 4.01 2.80 6.0 6.0 96.2 Totally disagree (5) 42 3.08 3.01 3.8 3.8 100 Total 1102 4.22 2.72 99.8 100 Missing System 2 .2 Total 1104 100
30 Table 4. Descriptive statistics of WTP-level (in euros) per Perceived Cause Importance dimension.
Perceived Cause Importance N Mean SD Percent Valid Percent
Cumulative Percent Valid Totally agree (1) 287 4.57 2.62 26.0 26.0 26.0
Agree (2) 364 4.38 2.62 33.0 33.0 59.1 Neutral (3) 331 4.02 2.77 30.0 30.0 89.1 Disagree (4) 77 4.02 2.84 7.0 7.0 96.1 Totally disagree (5) 43 2.55 2.94 3.9 3.9 100 Total 1102 4.22 2.72 99.8 100 Missing System 2 .2 Total 1104 100
4.2 Consumers’ purchase behavior
In this study a significance level of p < .05 was maintained for all conducted tests. In order to test Hypothesis 1, a one-way ANOVA was executed to compare the WTP-levels of consumers exposed to abstraction appeals and spatial distance appeals, and consumers exposed to neutral appeals (i.e. the different conditions), F(4, 1099) = 1.66, p = .156. The effect of appeals seemed not to be significant. This means that consumers exposed to abstraction appeals and spatial distance appeals were not willing to pay more for the product, than consumers exposed to neutral appeals. Thus, advertisements with abstraction and spatial distance appeals had no different effects on purchase behavior, than advertisements with neutral appeals. This finding was not in line with our first hypothesis.
Simultaneously, the results of this ANOVA can be used to test Hypotheses 3a and 3b. Hypothesis 3a compares consumers’ WTP-level within the correct fit conditions [Condition 1 & Condition 4], and Hypothesis 3b compares this within the incorrect fit conditions [Condition 2 & Condition 3]. The results above show that none of the conditions significantly differed from another. In case of the correct fit conditions: consumers exposed to concrete appeal with low spatial distance [Condition 1] were not willing to pay more for the product, than consumers exposed to abstract appeal with high spatial distance [Condition 4]. Thus, in
31 the advertisements with a correct fit in appeal level: concrete low spatial distance appeals had no different effects on purchase behavior, than abstract high spatial distance appeals. This finding did not support Hypothesis 3a. Nevertheless, the results of the same ANOVA support Hypothesis 3b, since we found earlier that the conditions did not differ significantly in their WTP-level. When applied to the incorrect fit conditions: consumers exposed to concrete appeal with high spatial distance [Condition 2] did not differ in their WTP-level, compared to consumers exposed to abstract appeal with low spatial distance [Condition 3]. Thus, in the advertisements with an incorrect fit in appeal level: concrete high spatial distance appeals had no different effects on purchase behavior, than abstract low spatial distance appeals. This finding was in line with our expectation.
In order to test Hypothesis 2 an additional one-way ANOVA was conducted to compare the WTP-levels of the correct fit conditions together [Condition 1 + Condition 4] and the incorrect fit conditions together [Condition 2 + Condition 3], F(1, 1102) = .26, p = .612. The effect of a correct fit in appeal level seemed not to be significant. This means that consumers in the correct fit conditions (N = 480, M = 4.27, SD = 2.7) were not willing to pay more for the product, than consumers in the incorrect fit conditions (N = 624, M = 4.19, SD = 2.74). Thus, advertisements with a correct fit in appeal level had no different effects on purchase behavior, than advertisements with an incorrect fit in appeal level. This finding did not support Hypothesis 2.
In order to test the moderating effects of Environmental Involvement (Hypothesis 4a) and Perceived Cause Importance (Hypothesis 4b), the “conditional process analysis” of Hayes was used (www.afhayes.com). This procedure is known as the PROCESS Procedure for Regression, whereby the letter (X) stands for the independent variable; (Y) for the dependent variable; and (M) for the moderating variable. For Hypothesis 4a, a PROCESS Regression (model 1: simple moderation) was conducted per condition to test the moderating influence of Environmental Involvement, see Table 5a till Table 5e. For most conditions, a
32 significant positive relation between Environmental Involvement and WTP-level was found: Condition 1, t(1098) = -2.87, p = .004; Condition 2, t(1098) = -3.34, p < .001; Condition 3,
t(1098) = -3.55, p < .001; and Condition 4, t(1098) = -2.55, p = .011. For Condition 5, this
was only marginally significant (at level p < .10), t(1098) = -1.74, p = .082. This means that consumers with a relatively high Environmental Involvement were willing to pay more for the product, than consumers with a relatively low Environmental Involvement. Thus, Environmental Involvement had a positive influence on purchase behavior, regardless of (the influences of) the advertisements’ appeals.
For Environmental Involvement in combination with abstraction and spatial distance appeals no significant interaction effects were found. Though, for Environmental Involvement in combination with neutral appeals [Condition 5] a significant positive moderating effect was found, t(1098) = -2.47, p = .014. This means that the effect of neutral appeals on consumers’ WTP-level depended on Environmental Involvement. The regression coefficient for the interaction (Neutral appeals x Environmental Involvement) is b3 = -.49. The interaction effect counts for 0.5% of variance in WTP-level, which is very small. Besides, the interaction effect of neutral appeals with Environmental Involvement seemed to be significant only when Environmental Involvement is higher than 3.8 (effect = -.80, SE = .41, CI: -1.598 to -.003), see Table 6 for the conditional effects. This can be interpreted as an interaction effect for low-involved consumers only in the relationship between neutral appeals and WTP-level. Beware of the scale dimensions of Environmental Involvement (1 = totally agree, 5 = totally disagree). Thus, Environmental Involvement positively moderated the relationship between the advertisement with neutral appeals and purchase behavior only. Even though we found a positive moderating influence of Environmental Involvement, this finding was not in line with Hypothesis 4a. We expected an interaction effect of Environmental Involvement in combination with abstraction and spatial distance appeals, and thus not in combination with neutral appeals.
33 Table 5a. Condition 1: Linear model of predictors of WTP-level.
C1 [Concrete, low spatial distance] Coefficient SE t p
Intercept il 4.68 .21 22.39 < .001 Condition 1 (X) b1 .29 .42 .70 .487 Environmental involvement (M) b2 -.26 .09 -2.87 .004 Environmental involvement x Condition 1 (XM) b3 .05 .18 .27 .786 Note. R2 = 0.013 F(3, 1098) = 4.68, p = .003
Table 5b. Condition 2: Linear model of predictors of WTP-level.
C2 [Concrete, high spatial distance] Coefficient SE t p
Intercept il 4.80 .20 24.01 <.001 Condition 2 (X) b1 -.40 .47 -.85 .397 Environmental involvement (M) b2 -.28 .09 -3.34 <.001 Environmental involvement x Condition 2 (XM) b3 .24 .20 1.16 .244 Note. R2 = 0.010 F(3, 1098) = 3.82, p = .010
Table 5c. Condition 3: Linear model of predictors of WTP-level.
C3 [Abstract, low spatial distance] Coefficient SE t p
Intercept il 4.92 .20 24.72 <.001 Condition 3 (X) b1 -1.04 .48 -2.18 .030 Environmental involvement (M) b2 -.30 .08 -3.55 <.001 Environmental involvement x Condition 3 (XM) b3 .32 .21 1.52 .128 Note. R2 = 0.014 F(3, 1098) = 5.20, p = .001
Table 5d. Condition 4: Linear model of predictors of WTP-level.
C4 [Abstract, high spatial distance] Coefficient SE t p
Intercept il 4.74 .20 23.51 <.001 Condition 4 (X) b1 -.08 .46 -.17 .862 Environmental involvement (M) b2 -.22 .09 -2.55 .011 Environmental involvement x Condition 4 (XM) b3 -.06 .19 -.34 .735 Note. R2 = 0.010 F(3, 1098) = 3.73, p = .011
34 Table 5e. Condition 5: Linear model of predictors of WTP-level.
C5 [Neutral] Coefficient SE t p Intercept il 4.52 .20 22.26 <.001 Condition 5 (X) b1 1.04 .45 2.34 .020 Environmental involvement (M) b2 -.15 .09 -1.74 .082 Environmental involvement x Condition 5 (XM) b3 -.49 .20 -2.47 .014 Note. R2 = 0.015 F(3, 1098) = 5.38, p = .001
Table 6. Conditional effect at the different levels of Environmental involvement in Condition 5 [Neutral].
Environmental involvement
Unstandardized Boot Effects
Boot SE Boot LLCI Boot ULCI Conditional effect at levels of Environmental involvement
Totally agree (1) .559 .287 -.004 1.122 Agree (2) .073 .206 -.331 .478 Neutral (3) -.412 .283 -.967 .142 Disagree (4) -.898 .441 -1.763 -.033 Totally disagree (5) -1.383 .621 -2.602 -.164
In order to test Hypothesis 4b, PROCESS Regression (model 1: simple moderation) was conducted per condition to test the moderating influence of Perceived Cause Importance, see Table 7a till Table 7e. In all conditions, a significant positive relation between Perceived Cause Importance and WTP-level was found: Condition 1, t(1098) = -3.72, p <.001; Condition 2, t(1098) = -4.72, p < .001; Condition 3, t(1098) = -4.16, p < .001; Condition 4,
t(1098) = -3.65, p < .001; and Condition 5, t(1098) = -3.27, p = .001. This means that
consumers with a relatively high Perceived Cause Importance were willing to pay more for the product, than consumers with a relatively low Perceive Cause Importance. Thus, Perceived Cause Importance had a positive influence on purchase behavior, regardless of (the influences of) the advertisement’s appeals.
For Perceived Cause Importance in combination with the advertisement appeals of Condition 1, 3, 4 and 5 no interaction effects were found. Conversely, for Perceived Cause
35 Importance in combination with the advertisement appeals of Condition 2 [Concrete, high spatial distance] a marginally significant positive moderating effect was found, t(1098) = 1.78, p = .076 (at level p < .10). This means that the effect of “concrete, high spatial distance” appeal on consumers’ WTP-level depended on Perceived Cause Importance. The regression coefficient for the interaction (Concrete, high spatial distance appeals x Perceived Cause Importance) is b3 = -.37. The interaction effect counts for 0.3% of variance in WTP-level, which is very small. Moreover, the interaction effect of “concrete, high spatial distance” appeal with Perceived Cause Importance seemed to be significant only when Perceived Cause Importance is higher than 3.4 (effect = .505, SE = .307, CI: .000 to 1.010), see Table 8 for the conditional effects. This can be interpreted as an interaction effect for consumers with a low Perceived Cause Importance only in the relationship between the particular appeals and WTP-level. Beware of the scale dimensions of Perceived Cause Importance (1 = totally agree, 5 = totally disagree). Thus, Perceived Cause Importance positively moderated the relationship between the advertisement with “concrete high spatial distance” appeals and purchase behavior only. This finding supported Hypothesis 4b partly, since we expected a positive moderating influence of Perceived Cause Importance (i.e. strengthening effect) with all the abstraction and spatial distance appeals. But, this strengthening effect took place only in combination with “concrete, high spatial distance” appeals.
Table 7a. Condition 1: Linear model of predictors of WTP-level.
C1 [Concrete, low spatial distance] Coefficient SE t p
Intercept il 4.89 .22 21.86 <.001
Condition 1 (X) b1 .46 .46 1.00 .318
Perceived Cause Importance (M) b2 -.33 .09 -3.72 <.001 Perceived Cause Importance x
Condition 1 (XM)
b3 -.03 .18 -.18 .859
36 Table 7b. Condition 2: Linear model of predictors of WTP-level.
C2 [Concrete, high spatial distance] Coefficient SE t p
Intercept il 5.12 .22 23.91 <.001
Condition 2 (X) b1 -.75 .52 -1.44 .150
Perceived Cause Importance (M) b2 -.40 .08 -4.72 <.001 Perceived Cause Importance x
Condition 2 (XM)
b3 .37 .21 1.78 .076
Note. R2 = 0.020 F(3, 1098) = 7.52, p < .001
Table 7c. Condition 3: Linear model of predictors of WTP-level.
C3 [Abstract, low spatial distance] Coefficient SE t p
Intercept il 5.12 .22 23.66 <.001
Condition 3 (X) b1 -.59 .51 -1.16 .246
Perceived Cause Importance (M) b2 -.36 .09 -4.16 <.001 Perceived Cause Importance x
Condition 3 (XM)
b3 .10 .20 .48 .628
Note. R2 = 0.020 F(3, 1098) = 7.46, p < .001
Table 7d. Condition 4: Linear model of predictors of WTP-level.
C4 [Abstract, high spatial distance] Coefficient SE t p
Intercept il 5.00 .22 23.09 <.001
Condition 4 (X) b1 .01 .50 .03 .978
Perceived Cause Importance (M) b2 -.31 .09 -3.65 <.001 Perceived Cause Importance x
Condition 4 (XM)
b3 -.11 .20 -.54 .589
Note. R2 = 0.019 F(3, 1098) = 6.94, p < .001
Table 7e. Condition 5: Linear model of predictors of WTP-level.
C5 [Neutral] Coefficient SE t p
Intercept il 4.87 .22 21.94 <.001
Condition 5 (X) b1 .60 .47 1.28 .202
Perceived Cause Importance (M) b2 -.29 .09 -3.27 .001 Perceived Cause Importance x
Condition 5 (XM)
b3 -.24 .19 -1.27 .204
37 Table 8. Condition 2: Conditional effect at the different levels of Perceived Cause Importance.
Environmental involvement
Unstandardized Boot Effects
Boot SE Boot LLCI Boot ULCI
Conditional effect at levels of Perceived Cause Importance
Totally agree (1) -.380 .340 -.940 .181 Agree (2) -.011 .216 -.366 .345 Neutral (3) .359 .252 -.056 .775 Disagree (4) .729 .408 .056 1.400 Totally disagree (5) 1.098 .597 .116 2.080
As announced before, we decided to analyze the influence of Familiarity, Liking (of the product) and some Demographics on consumers’ WTP-level, despite we did not state hypotheses for these concepts. Familiarity. An independent t-test was conducted to compare the WTP-levels of familiar and unfamiliar consumers, t(1099) = -7.25, p < .001. The results show that consumers who were familiar with the product (N = 817, M = 4.57, SD = 2.60) were willing to pay significantly more than consumers who were unfamiliar with the product (N = 284, M = 3.24, SD = 2.81). Thus, this finding indicates that Familiarity with the product had a positive influence on purchase behavior.
Liking. The descriptive statistics of Liking are presented in Table 9. OLS was
conducted to predict WTP-level based on Liking. The regression equation was found to be significant (F(1, 1102) = 734.07, p < .001, R2 = .090). The predicted consumers’ WTP-level is equal to 5.729 - .594 (Liking) euros when Liking is measured in (five) dimensions. The WTP-level decreased .594 (euros) for each dimension of Liking. Beware of the scale dimensions of Liking (1 = totally positive, 5 = totally negative). Thus, this finding indicates that Liking predicted purchase behavior to a moderate extent.
Demographics. When investigating some demographic variables, only significant
effects were found for Gender, Household size and Age. (1) Gender: an independent t-test was executed to compare the WTP-levels of men and women, t(1102) = -5.83, p < .001. The
38 results show that women (N = 554, M = 4.69, SD = 2.50) were willing to pay significantly more for the product compared to men (N = 550, M = 3.75, SD = 2.85). Thus, this finding indicates that women’s purchase behavior was influenced more positively than men’s purchase behavior. (2) Household size: the descriptive statistics are presented in Table 10. OLS was conducted to predict WTP-level based on Household sizes. The regression equation was found to be significant (F(1, 656) = 13.24, p < .001, R2 = .020). The predicted consumers’ WTP-level is equal to 3.227 + .334 (Household size) when Household size is measured in the amount of persons. The WTP-level increased .334 (euros) for each person in Household. Thus, this finding indicates that Household size predicted purchase behavior to a moderate extent. (3) Age: OLS was conducted to predict WTP-level based on Age. The regression equation was found to be significant (F(1, 1102) = 10.53, p = .001, R2 = .009). The predicted consumers’ WTP-level is equal to 5.004 - .018 (Age) when Age is measured in years. The WTP-level decreased for .018 (euros) per Age year. Thus, this finding indicates that Age predicted purchase behavior to a small extent.
Table 9. Descriptive statistics of WTP-level(in euros) per Liking dimension.
Liking N Mean SD Valid Percent Cumulative Percent
Totally positive (1) 250 4.98 2.40 22.6 22.6 Positive (2) 378 4.93 2.41 34.2 56.9 Neutral (3) 411 3.42 2.81 37.2 94.1 Negative (4) 47 2.74 2.83 4.3 98.4 Totally negative (5) 18 1.14 2.20 1.6 100 Total 1104 4.22 2.72 100
