Master Thesis
Collaborative consumption for a
sustainable future: What gets
consumers on-board?
An investigation of the consumers' willingness to
adopt PSS and its determinants
Author: Wessman, Anna Supervisor: Sattari, Setayesh Examiner: Pehrsson, Anders Date: 11/02/2019
Subject: Business Administration
with specialization in Innovation
Abstract
The Circular Business Model (CBM) is one way to achieve a more sustainable future. Its goal is to reduce the use of natural resources which are becoming more scarce on our planet. The aim of the study is to investigate the consumer behavior and the determinants of the willingness to adopt Product Service System (PSS), one of the models of CBM. Through previous research, five main factors were identified: Economical Value, Flexibility, Trust, Desire to Own and Peer Influence. Furthermore, some literature argue that environmental awareness has an influence on the consumers’ choice of consumption. Therefore, this factor is considered and used as a moderating variable in the conceptual model. The primary data was collected through an online questionnaire and the hypotheses were tested through multiple regression analyses. Seven out of the twelve hypotheses were accepted. The study also concludes that environmental awareness does not have a high impact on the model. The environmental awareness did only have a positive moderating effect in the peer influence factor. It was determined in the study that economical value has the highest impact on the model.
Keywords
Circular Business Models, CBM, Product Service System, PSS, Willingness to Adopt, Sustainability, Consumer Behavior, Environmental Awareness, Economics
Acknowledgement
Contents
1 Introduction ________________________________________________________ 5 1.1 Background ______________________________________________________ 5 1.2 Problem Discussion ________________________________________________ 7 1.3 Purpose _________________________________________________________ 9 2 Theoretical framework ______________________________________________ 10 2.1 Introduction to PSS _______________________________________________ 10 2.2 Willingness to adopt PSS __________________________________________ 10 2.3 Determinants of PSS ______________________________________________ 11 2.3.1 Economical Value ____________________________________________ 11 2.3.2 Flexibility ___________________________________________________ 12 2.3.3 Trust _______________________________________________________ 13 2.3.4 Desire to Own _______________________________________________ 14 2.3.5 Peer Influence _______________________________________________ 142.4 Moderating Role of Environmental Awareness _________________________ 15 2.5 Conceptual Framework and Model ___________________________________ 17
3 Method ___________________________________________________________ 19
3.1 Research Approach _______________________________________________ 19 3.2 Research Design _________________________________________________ 19 3.3 Data Sources and Sampling ________________________________________ 20 3.4 Data Collection __________________________________________________ 21
3.4.1 Pre-study ___________________________________________________ 22
3.5 Data Collection Instrument _________________________________________ 22
3.5.1 Operationalization ____________________________________________ 22 3.5.2 Pretesting ___________________________________________________ 25
3.6 Data Analysis Method _____________________________________________ 25
3.6.1 Quality Criteria ______________________________________________ 25 3.6.2 Hypothesis Testing ____________________________________________ 26
3.7 Ethical Considerations ____________________________________________ 27
4 Data Analysis ______________________________________________________ 29
4.1 Descriptive Statistics ______________________________________________ 29 4.2 Validity and Reliability Analysis ____________________________________ 29 4.3 Hypothesis Testing _______________________________________________ 31 4.4 Additional Findings _______________________________________________ 33
5 Discussion _________________________________________________________ 34
6 Conclusions and Contributions ________________________________________ 40
6.1 Theoretical Contributions __________________________________________ 40 6.2 Managerial Implications ___________________________________________ 41 6.3 Limitations and Suggestions for Further Research _______________________ 41
References __________________________________________________________ 43 Appendices ___________________________________________________________ I
List of Tables
Table 1 - Summary of Theoretical Framework ... 17
Table 2 - Operationalization of Theoretical Concepts ... 23
Table 3 - Premliminary Rotated Component Matrix ... 30
Table 4 - Reliability Testing ... 31
Table 5 - Regression Analysis Without Moderation ... 32
Table 6 - Summary of the Regression Analysis ... 32
Table 7 - Age of Respondents ... XII Table 8 - Gender of Respondents ... XII Table 9 - Occupation of Respondents ... XII Table 10 - Type of Consumer of Respondents ... XII Table 11 - Rotated Component Matrix of Flex and Trust ... XIII Table 12 - Correlation Analysis of Flexibility and Trust ... XIII Table 13 - Descriptive Analysis ... XIV Table 14 - Pearson’s r Correlation Test of the Independent Variables ... XIV Table 15 - Regression Analysis with Moderation on Economic Value ... XV Table 16 - Regression Analysis with Moderation on Flexibility ... XV Table 17 - Regression Analysis with Moderation on Trust ... XVI Table 18 - Regression Analysis with Moderation on Desire to Own ... XVI Table 19 - Regression Analysis with Moderation on Peer 1 ... XVII Table 20 - Regression Analysis with Moderation on Peer 2 ... XVII Table 21 - ANOVA Test ... XVIII
List of Figures
Figure 1 - Conceptual Model ... 181 Introduction
In this introduction chapter the background to the study is presented followed by the problem discussion and purpose of the study.
1.1 Background
The constantly growing economic system, developed by humans, puts an enormous amount of stress on the planet’s resources. The pressure on the world’s natural resources and a growing population creates challenges for a sustainable future which can be approached in different ways. Before, the main focus has been on efficiency and productivity improvements (Bocken and Short, 2016). However, the current growth is by far outpacing the innovations and improvements resulting in even more consumption (Druckerman et al. 2011; Bocken et al. 2014). This has led to environmental concerns on a global level and a realization of the need to develop new ways to conduct business and new business models (Mathews, 2011; Lewandowski, 2016).
The most common way to conduct business today is to use a linear business model, a model that put further stress on the planet. The linear model (LBM) is a “take-make-use-dispose” situation where resources are seen as a means to sell artifacts to a profit. Whether or not these artifacts are recyclable or not is generally not considered in this business model (Bakker et al. 2014). This way of thinking and doing business has a huge negative impact on the environment and it calls for a change in the way companies conduct their business in order to become sustainable (Ellen McArthur Foundation, 2013; Bakker et al. 2014).
industry remanufacturing parts, or electronic stores offering cash for electronics, refurbishing and selling them to another customer (Bocken et al. 2016).
To implement CBM, Linder and Willander (2017) argue that there are several challenges and that they differ depending on the businesses, products and customer segments. However, within a business, the model needs to be looked upon by both the supply and the demand side. One needs full collaboration from all stakeholders and partners along the supply chain (Van Bruen et al. 2016). Studies have been done looking into the supply side and the implementation. In Linder and Willander’s (2017) study, they found that there is always a higher business risk within a CBM compared to a corresponding LBM. These risks could be difficult to overcome since the business partners along the supply chain might not be willing to prioritize a CBM in their business (Luthra et al. 2011). Adopting CBM in the supply chain is complex and it raises questions of the value chain in terms of ownership, costs, resources and benefits (Rizos et al. 2016).
Another challenge for CBM is the consumer and the demand side. There are several issues regarding the consumers, and one of them is the fact that the consumers do not have sufficient awareness of the benefits of a CBM, or in their meaning, green consumption (Zhu et al. 2008; Wooi and Zailani, 2010; Edbring, 2016). All the way back to the beginning of the industrialization, the consumers in Western society have been taught the throwaway mentality (Cooper, 2013). Watson (2008) argues that products today are not designed to last or to be repaired. Products are rather bought, swiftly consumed and then discarded. The transition to CBM necessitates a change in the consumer behavior and a change in the way the consumers look at consumption and ownership of products and resources (Edbring et al. 2016). Rizos et al. (2016) argue that the companies cannot go through this transition without a change in both the value chain and the consumers. When going through a transition, it is therefore important that all different stakeholders involved are on board and ready to incorporate the new structure. In the transition from a linear business model to a circular one, every part is affected – everything from the raw material extraction to the consumer and recycling unit.
In 2017, the United Nations University (UNU) conducted a study where they monitored the world’s e-waste, which includes everything from white goods, screens, phones, and lamps to USB sticks. It showed that in 2016, the world produced 44.7 million tons of e-waste with an estimated value of the raw material of 55 billion euros (Baldé et al. 2017). It forecasted that in 2022 the amount of e-waste would be as high as 52.2 million ton with an average of 6.8 kg/inh (Baldé et al. 2017). In Europe, the average e-waste per person was as high as 16.6 kg/inh in 2016 and only 4,3 kg is documented as recycled (Baldé et al. 2017). With this in mind, it is not hard to see that the consumers have no issue spending money on home electronics. In 2016, the ten leading consumer electronics retailers in Europe had a turnover of 78.7 billion euros and the market is expected to grow with 2% in the upcoming years (Statista, 2018a; Statista, 2018b). This makes this sector of consumer goods fitting to investigate both from a sustainable perspective and an economical perspective.
1.2 Problem Discussion
One approach, that is gaining more momentum within CBM, is for the company to deliver a function rather than a product (Bocken et al. 2014). This is called access-based consumption or Product Service System (PSS), which emphasizes the use and functions of the product rather than owning the product (Catulli, 2012; Edbring et al. 2016). The main focus is not on the tangible product, but on the customer experience and the function the product provides (Catulli, 2012; Edbring et al. 2016). In this situation, compared to the situation above where legislators regulate the market, the link between profit and production volume is broken without reducing usage volume (Bocken et al. 2014). Examples of this scenario could be car sharing, furniture rental, or leasing jeans (Bocken et al. 2016). That means that you could either have a long-term leasing deal or rental agreement or just a short-term solution based on the actual consumption.
Since the consumers do not, in this CBM scenario, own the product but rather buy the service it provides, the manufacturer needs to take more responsibility. The manufacturer needs to close the material loops which is the core of CBM (Ellen McArthur Foundation, 2013; Mentink, 2014; Lewandowski, 2016). Companies and manufacturers adopting CBM need to put more emphasis on product durability and how to use energy and resources more efficiently, since they retain the ownership of the asset throughout the whole life cycle of the product (Mont, 2008; Bocken et al, 2014; Rizos et al. 2016). This means that the companies need to get to know and understand the consumers in a new way. The companies need to know and understand how the consumers would adopt this service rather than buying a product (Lewandowski, 2016).
that the consumers’ willingness to adopt this business model is one main concern of CBM and the companies heading towards this future. Since the consumer is one of the main stakeholders of the model the companies need to understand how to get the consumers involved. Studies have shown that the consumers do not have sufficient awareness regarding CBM (Zhu et al. 2008; Wooi and Zailani, 2010; Edbring, 2016) and that there is a lack of information and understanding regarding the concept (Catulli, 2012; Rizos et al. 2016). Roos (2014) and Lewandowski (2016) argue that it is not easy to get the consumers to change their behaviors and habits to a CBM behavior and a more function-based consumption. Considering this, Catulli (2012) further argues that one main issue is the concern the consumers have when it comes to satisfaction; if the service will live up to their expectations or not. Thus, there is a risk that the consumers become reluctant to further explore the opportunities. As Lewandowski (2016) states, there is a lack of information for the companies regarding the consumers’ pains and gains related to the CBM. Therefore, it is important to look at CBM from a consumer adoption perspective to better understand and know what they consider important.
To gain more knowledge and to understand what makes the consumers willing to adopt and pay for a product or service in relation to environmental benefits, accompanying CBM, there are different factors that could be studied. As argued by Bocken et al. (2014), the consumers do not want to decrease their consumption and their usage of home electronics. However, what can make them willing to agree to a PSS situation where they do not own their furniture, but rather lease it during a period of time? Leasing and renting models rely on the retained value and should reduce material costs which is important for the company (Vezzoli et al., 2015). However, they are often difficult to implement since it is a fundamental change in the interaction with the consumers, (Mont et al. 2006; Tukker, 2015) and it does challenge the consumers preferences (Besch, 2005; Armstrong et al. 2015).
and barrier of PSS (Edbring et al. 2016). Further, peer influence has been recognized as one of the primary factors influencing consumer decisions (Churchill and Moschis, 1979; Mangleburg et al. 2004; Luo, 2005; Baek and Ho, 2015).
All the previous mentioned factors have, according to previous research, an impact on the consumers’ willingness to adopt. Although, there are some concerns when it comes to the actual impact of environmental awareness. Liu et al. (2017) conclude that environmental awareness is not always consistent, meaning that there is a gap between the knowledge and the actual behavior. Peattie (2010) found that environmental knowledge and awareness is a key factor in sustainable consumption. However, Bartiaux (2008) shows that there is no causality between environmental knowledge and action. This raises the question of what impact the environmental awareness has on the willingness to adopt a PSS situation and how that relates to the other factors identified above. There are uncertainties involving the consumers that relate to CBM expressed by several authors (Zhu et al. 2008; Wooi and Zailani, 2010; Catulli, 2012; Edbring et al. 2016). To investigate this further could lead to more knowledge about whether the consumers would be willing to adopt the consumer behavior that the PSS situation requires or not.
1.3 Purpose
2 Theoretical framework
In this chapter, the literature review of each identified factor in the problem discussion is presented. The literature review starts with a brief review of PSS and then continues with the willingness to adopt PSS. Then it moves forward with the other determinants of the willingness to adopt PSS. Lastly, the moderating variable of environmental awareness is discussed followed by the conceptual framework and the conceptual model.
2.1 Introduction to PSS
Product Service System (PSS) is a new type of business model which is a part of CBM. PSS is a type of access-based consumption that emphasizes the use and functions of the product instead of owning the product (Catulli, 2012; Edbring et al. 2016). The main focus is not on the tangible product but on the customer experience and the function it provides (Catulli, 2012; Edbring et al. 2016). PSS models are emerging as a collaborative consumption platform of both products and services, where consumers can rent or borrow products with the aim of pro-environmental outcomes – such as reducing the use of natural resources (Piscicelli et al. 2014). PSS can be seen as a result-oriented service that provides access to products that is still owned and maintained by the company (Piscicelli et al. 2014).
2.2 Willingness to adopt PSS
When discussing the willingness to adopt, one needs to consider that there are several different stages of adoption. The most common model to use when discussing adoption theory is Everett Rogers’ model of diffusion of innovation (Solomon et al. 2016; Palmatier and Sridhar, 2017). To understand what has an impact of the consumers’ willingness to adopt, there are several aspects to consider within Rogers’ theory of adoption. There are four stages of the innovation-decision process. These are knowledge, persuasion, implementation and confirmation (Rogers, 2003). In the knowledge phase, the consumers get exposed to the existence of the product or service and gain understanding about it (Rogers, 2003; Sahin, 2006). However, it is in the next stage, the persuasion phase, that the willingness of the adoption process is tested (Rogers, 2003; Sahin, 2006). Sahin (2006) argues that it is during the persuasion stage that the consumers weigh the perceived characteristics or attributes of the innovation, which affect the willingness. The implementation and confirmation phases focus more on the actual adoption when the consumers have evaluated the options (Rogers, 2003).
decision process is a favorable or unfavorable attitude towards the innovation (Rogers, 2003).
To investigate different factors of a new service is important in order to know how well it will be adopted. Rogers have identified five general characteristics that explain 49% to 87% of the variance of the willingness to adopt (Rogers, 2003). These are relative advantage, compatibility, complexity, trialability, and observability (Rogers, 1995). These measurements are the foundation of the adoptability and is therefore the measurements used in this study to decide the adoptability of PSS.
Other than Rogers stages of diffusion of innovation there are also different type of consumers. When a new product or service enters the market, different people have different adoption rates. According to Holak et al. 1987; Armstrong et al. 2012 there are three main categories, early adopters, majority (early/late), and laggards. Early adopters are among the first to try a new product and are in some cases seen as an opinion leader (Solomon et al. 2013). After the early adopters, the majority adopts the product, this group of people can sometime wait deliberately before adopting an innovation due to price or expected improvement on the product or service (Holak et al. 1987). Finally, laggards are the last to adopt new things, this because they are very tradition bound and do not like change (Armstrong et al. 2012)
2.3 Determinants of PSS
In the following chapters, each of the determinants of PSS is discussed. These determinants were identified in the problem discussion and is further discussed here.
2.3.1 Economical Value
However, as argued, cost is also related to behavior. Tobler et al. (2012) argue that there are several factors that has an impact on our cost behavior when it comes to environmental benefit. These can be time, discomfort, effort and other subjectively defined inhibitions of one’s behavior (Diekmann and Preisendörfer, 1998; Kollmuss and Agyeman, 2002; Tobler et al. 2012). This means that the consumers compare the behavioral costs and efforts that need to be made in order to adapt and conform to the environmental benefit (Tobler et al. 2012). According to Tobler et al. (2012), price in itself is usually not enough to change the perception of behavioral cost. Barber et al. (2016) argue that the consumers need another incentive to change their behavior, such as loyalty bonuses or other benefits. Further, Diekmann and Preisendörfer (2003) argue that when the cost is low the transformation to the desired behavior is not difficult to achieve due to environmental benefit. However, when the consumers feel that the behavioral cost is too great, environmental concerns are not sufficient to overcome the barriers.
As one can understand, the economic value of a service is of outmost importance for a potential consumer (Tobler et al. 2012; Edbring et al. 2016). It has an impact on whether the consumer chooses to be a part of the arrangement or not. In this context, the economic value is argued have a positive impact on the consumer in the conceptual model which is described in chapter 2.5. To understand one’s behavior and recognize the perceived monetary value of a service that accommodates environmental benefit is a challenge that needs to be investigated (Tobler et al. 2012; Liu et al. 2017).
H1: Economical value of PSS has a positive impact on the consumers’
willingness to adopt PSS
2.3.2 Flexibility
opportunity and variability has been proven to be a strong motivator in products that require maintenance and upgrade, for example products with fast innovation cycles (Rexefelt and Hjort af Ornäs 2009; Edbring et al. 2016). In those scenarios, the attractiveness of renting is higher.
Flexibility is a factor that is seen as a strong positive aspect of PSS (Rexefelt and Hjort af Ornäs 2009; Edbring et al. 2016). To increase the variability for the consumers and to have a flexible arrangement are important factors for PSS in order to be attractive; therefore, it will be treated as a factor with positive impact on the willingness to adopt a PSS behavior.
H2: Flexibility of PSS has a positive impact on the consumers’ willingness
to adopt PSS
2.3.3 Trust
In most current literature, trust and risk are two concepts that are often linked together (Ding et al. 2013). Some argue that trust comes into play only in situations involving risk and uncertainty (Boon and Holmes, 1991; Ding et al. 2013). When this occurs, trust may compensate for missing information by strengthening an individual’s feelings of confidence and internal security: therefore, trust helps reducing the difficulty of making a decision involving risk (Luhmann, 1979).
When agreeing to a PSS, there is a relationship between the consumer and the provider, which comes with some attachments. To form positive attitudes towards a PSS provider, the consumer needs to trust the provider (Catulli et al. 2013; and Armstrong et al. 2015). Since PSS is a rather new way to consume, there is a lot of perceived risk in the consumer’s mind (Edbring et al. 2016). However, Baumeister (2014) argues that this perception decreases as the consumer gets more accustomed to PSS and start to trust the relationship. Studies have shown that some consumers feel that, when using rented products in a personal manner, the company is interfering in one’s personal space (Raja et al. 2013; Edbring et al. 2016). This means there needs to be a lot of emphasis on the openness and trust between the provider and consumer to facilitate a good relationship (Raja et al. 2013; Edbring et al. 2016).
To have a high sense of trust in the PSS arrangement and the company providing the service is important and will have a positive impact on the willingness to adopt the PSS concept (Catulli et al. 2013; Armstrong et al. 2015). If one were to measure the perception of risk, it would have a negative impact on the model, however, in the scenario that is going to be investigated, one assumes that the provider can maintain a high sense of trust towards the consumer.
H3: Trust in PSS has a positive impact on the consumers’ willingness to
2.3.4 Desire to Own
Ever since the industrial revolution, product ownership has been one of the main characteristics of the modern consumption culture (Wallendorf and Arnould, 1988). This means that there is an institutionalized social norm which has taught generations of the importance of ownership. According to Solomon et al. (2007), product ownership is a way for persons to project an image of themselves as well as satisfying a need for a product’s function. Catulli (2012) and Edbring et al. (2016) both concluded that ownership is an inherent value in the modern consumption society and that it has an important impact on how people view products and how they treat leased products. Edbring et al. (2016) found that the fear of breaking rented products is an obstacle of PSS. People are uneasy about using rented products in the same way as they would with products they own since they feel that the products need to be handled with more care. They also found that consumers have a fear of health risks and sanitation issues associated with renting product that have previously been used by an unknown person.
Ebring et al. (2016) found that the lack of ownership seems to negatively influence attitudes towards PSS. They found that people were rather negative towards not owning the products that they use for a long period of time. The consumers then believe that it would be easier and less expensive to buy the product. However, when it comes to short term rental agreements, the PSS is more accepted and the desire to own is not as high. Further, Bardhi and Eckhardt (2012) show that when people use a PSS in a short-term agreement they do not develop the sense of ownership over the product because of the temporary involvement with the product.
In this study, the factor of desire to own has a negative direction. As Catulli (2012) and Edbring et al. (2016) argue, ownership is an inherent value in the modern consumption society. This means that the consumers look upon ownership as the normal behavior when consuming, and that will consequently be an obstacle for them to overcome in order to participate in a PSS environment. This factor is treated as having a negative influence on the model, just as Edbring et al. (2016) argue that it will have a negative impact on the consumers’ attitudes.
H4: Desire to own has a negative impact on the consumers’ willingness to
adopt PSS
2.3.5 Peer Influence
people might use peer influence as a reason to justify their own consumption, and that peer consumption might also encourage them to consume in a certain way; for example, trying PSS for the first time.
Peer influence and peer consumption are highly related to hedonic purchase decisions (Mangleburg et al. 2004; Beak and Ho, 2015). Hedonic consumption is intangible and more related to the consumers’ expressive needs (Beak and Ho, 2015). It is through hedonic consumption people communicate with others and display their personal tastes (Hirschman and Holbrook, 1982) and therefore, it is influenced by the peers around them (Beak and Ho, 2015). Further, Mangleburg et al. (2004) argue that peers can influence someone to spend more money than they would originally do, and that the mere presence of peers can enhance the consumers’ attitudes toward a company.
Based on the literature, there is a higher chance to get the consumers to participate in a PSS if their peers are doing it. As stated by several authors, peer influence is one of the main factors influencing consumer behavior. Therefore, one can assume that if several peers are considering or participating in the arrangement, the chances are higher that a person will. Based on this, peer influence will be treated as having a positive impact on the willingness to adopt.
H5: Peer influence has a positive impact on the consumers’ willingness to
adopt PSS
2.4 Moderating Role of Environmental Awareness
Although there are several authors that argue for the importance of an environmental awareness on the consumer behavior, there are as many contradicting arguments towards this (Kikuchi-Uehara et al. 2016; Liu et al. 2017). Bartiaux (2008) argues that there is no connection between environmental knowledge and action. Liu et al. (2017) argue that the factor of environmental awareness is not always consistent. Both Kikuchi-Uehara et al. (2016) and Liu et al. (2017) argue that there is a gap between a consumer’s environmental awareness and the actual behavior. Kikuchi-Uehara et al. (2016) further argue that even if a consumer would have an increase in environmental awareness, it does not necessarily lead to a change in their behavior. Further, Zhao and Zhong (2015) argue that it is not environmental awareness that decides if a consumer chooses to have a pro-environmental consumer behavior, but rather the customer loyalty, the price, and the educational level. When considering the price and the premium one might have to pay in order to adopt a pro-environmental purchase behavior, Hazen et al. (2012), argue that environmental awareness is not enough. It is rather the customer loyalty that decides the behavior. There is a difference of opinion regarding how environmental awareness will impact the willingness to adopt a consumer behavior. There are arguments that support the idea that environmental awareness and a positive environmental attitude will lead to a pro-environmental consumer behavior (Gadenne et al. 2011; Assarut and Srisuphaolarn 2012; Arslan et al. 2012). However, some studies show that there is no relationship between the environmental awareness and the actual behavior (Kikuchi-Uehara et al. 2016; and Liu et al. 2017). Therefore, in this study environmental awareness is treated as a moderating variable in the conceptual model, moderating a positive relationship between the variables.
H6: Environmental awareness moderates the relation between economical
value of PSS and the consumers’ willingness to adopt PSS
H7: Environmental awareness moderates the relation between flexibility of
PSS and the consumers’ willingness to adopt PSS
H8: Environmental awareness moderates the relation between trust of PSS
and the consumers’ willingness to adopt PSS
H9: Environmental awareness moderates the relation between desire to
own of PSS and the consumers’ willingness to adopt PSS
H10: Environmental awareness moderates the relation between peer
2.5 Conceptual Framework and Model
The literature review above reveals several different factors that relate to the consumers’ willingness to adopt PSS. Within each factor, different measurements were identified. In table 1, each measurement is presented with its source in order to gain an overall understanding of the factor.
Table 1 - Summary of Theoretical Framework
Factor Measurement Reference
Economical value
Price Edbring et al. (2016), Liu et al. (2017)
Value Liu et al. (2017)
Product benefit Barber et al. (2016)
Time Diekmann and Preisendörfer (1998), Kollmuss and Agyeman (2002), Tobler et al. (2012) Discomfort Diekmann and Preisendörfer (1998), Kollmuss and Agyeman (2002), Tobler et al. (2012) Effort Diekmann and Preisendörfer (1998), Kollmuss and Agyeman (2002), Tobler et al. (2012)
Flexibility
Freedom Catulli et al. (2013), Edbring et al. (2016)
Commitment Rexefelt and Hjort af Ornäs (2009)
Variability Rexefelt and Hjort af Ornäs (2009), Edbring et al. (2016)
Guaranteed access Edbring et al. (2016)
Flexibility Rexefelt and Hjort af Ornäs (2009), Edbring et al. (2016)
Trust
Relationship Baumeister (2014) Raja et al. (2013), Edbring et al. 2016
Trust Edbring et al. (2016)
Openness Raja et al. (2013), Edbring et al. (2016)
Information Luhmann (1979)
Desire to own
Importance of ownership Wallendorf and Arnould (1988)
Self-image Solomon et al. (2007)
Handel with care Edbring et al. (2016)
Sanitation Edbring et al. (2016)
Peer influence
Guidance Churchill and Moschis, (1979) Mangleburg et al. (2004) Luo, (2005), Baek and Ho (2015)
Justification Beak and Ho (2015)
Social Identity Hirschman and Holbrook, (1982), Beak and Ho, (2015)
Expressive needs Mangleburg et al. (2004), Beak and Ho (2015)
Willingness to adopt
Advantage Rogers (1995), Rogers (2003)
Compatibility Rogers (1995), Rogers (2003)
Complexity Rogers (1995), Rogers (2003)
Trialability Rogers (1995), Rogers (2003)
Observability Rogers (1995), Rogers (2003)
Environmental awareness
Pro-environmental beliefs Gadenne et al. (2011), Liu et al. (2017)
Willingness to pay Birgelen et al. (2009), Dono et al. (2010), Gadenne et al. (2011) Environmental
The ten hypotheses presented in the literature review are shown in the model below. The independent variables of cost, flexibility, trust, desire to own and peer influence are tested towards the dependent variable of willingness to adopt PSS. These are followed by the hypotheses for each independent variable including the moderator: environmental awareness.
3 Method
This chapter presents and describes the methodology utilized as a blueprint to the research and the necessary considerations made prior to the study. It starts with the research approach and design and is followed by the sampling and data collection. Finally, the operationalization is presented along with the data analysis methods.
3.1 Research Approach
Bryman and Bell (2011) argue that there are two main types of research approaches: inductive and deductive. This study holds a deductive research approach, where the gathered theory from previous literature is the foundation of the whole research and where the hypothesis is derived and tested from that theory (Malhotra and Birks, 2007; Bryman and Bell, 2011). The research starts in the existing theory of CBM and PSS and investigates these topics with the aspiration to add to the existing knowledge (Malhotra and Birks, 2007; Bryman and Bell, 2011). According to Hyde (2000), the objective of a deductive research is to seek falsification of the initial finding through a search for support for an alternative explanation.
In this study, a deductive approach was applied to examine CBM and PSS, using the context of home electronics. The hypothesis is derived from the existing theory and is tested to be either accepted or rejected. In order to do this, a quantitative method is used. In research, there are two main types of research, qualitative and quantitative (Bryman and Bell, 2011), both with the purpose of connecting theory with social reality (Olsson and Sörensen, 2011) In quantitative research, the focus is on the generalizability and on the analysis of a big amount of data (Bryman and Bell, 2011; Creswell, 2003).
As argued above, the main part of this study has a quantitative research approach, where the aim is to collect quantitative data and then verify it and generalize it onto the population from where the sample is drawn (Creswell, 2003; Ghauri and Grønhaug, 2005). However, in the pre-study, a focus group was conducted in order to gain understanding of and insight into the consumers’ perspective, which is the aim of a qualitative study (Creswell, 2003; Bryman and Bell, 2011).
3.2 Research Design
According to Malhotra (2010), the research design is the framework and blueprint of research. He further argues that there are two main research designs: exploratory and conclusive. This study holds a conclusive research design where the aim is to test specific hypotheses and to examine the relationships between the variables (Malhotra, 2010). A conclusive research design typically has a large representative sample and the data is subjected to a quantitative analysis (Malhotra, 2010), which corresponds with the outline of this research. However, even if it is called conclusive research it is noted that within research, nothing can be proven and nothing is conclusive (Malhotra, 2010).
a descriptive and cross-sectional research design. These are most frequently used together (Malhotra, 2010). A descriptive research design is utilized when the aim is to describe a phenomenon (Saunders et al. 2009; Malhotra, 2010). According to Malhotra (2010), a descriptive research design requires a clear specification of the who, what, when, where and why, which corresponds with the foundation of this research. The cross-sectional design involves the data collection. The data is gathered from the sample population at one point in time and only once, compared to a longitudinal study where the data collection is done over time (Saunders et al. 2009; Malhotra, 2010; Bryman and Bell, 2011).
3.3 Data Sources and Sampling
When selecting data sources, one needs to carefully consider which sources to use, since they have a great impact on the research. There are two main types of data sources: primary and secondary (Malhotra and Birks, 2007; Fowler, 2014). The difference is that the primary data is collected by the researcher for the specific study, whereas the secondary data already has been collected for the purpose of others in another research or investigation (Malhotra and Birks, 2007). In this research, both primary and secondary data are used. According to Malhotra and Birks (2007), it is important to read and analyze secondary data regarding the studied phenomenon. They argue that it is an essential component of a successful research design. This corresponds well with the deductive approach of this study, where the literature review has built up the framework for the study and the collection of the primary data. The primary data was collected after a careful analysis of the secondary data in order to know what information was needed. Later, it was analyzed against the secondary data to draw conclusions and gain further knowledge within the field of the study.
Since this study involves collecting primary data, the data needs to derive from a sample, which are the people drawn from the target population (Fowler, 2014). This is a quantitative research, which means that one wants to reach a large sample where it is important that it is representative for the population and non-biased (Bryman and Bell, 2011; Heerwegh et al. 2005). Further, there are two main types of sampling: non-probability sampling and non-probability sampling (Ghauri and Grønhaug, 2010; Zikmund et al. 2010; Bryman and Bell, 2011). The difference between these samples is that in a probability sampling, every unit of the population has the same chance of being chosen, whereas in non-probability, sampling is not random (Ghauri and Grønhaug, 2010; Zikmund et al. 2010; Bryman and Bell, 2011).
the population (Bryman and Bell, 2011). When using these types of sampling techniques, the researcher needs to be aware of the issues that comes with it. As Bryman and Bell (2011) point out, these samples might be hard to generalize. This is because the sample might not be representative in the same manner as a probability sample would be. The researcher is aware of this and discusses it further in the reflection.
3.4 Data Collection
Data collection is the process where the researcher measures and gathers information in order to gain knowledge and information about the research topic (Saunders et al. 2009; Malhotra, 2010). Fowler (2014) argues that the data collection and its method is one of the most extensive choices the researcher has to make while preparing the study. For the research in this study, primary data was used and it was collected through a questionnaire. This was considered the most applicable method for fulfilling the purpose and the nature of this study. Malhotra (2010) argues that a questionnaire is a form of structured data collection where the questions are set in a pre-arranged order and designed and distributed to the sample in order to extract specific information from the respondents. Since questionnaires enable the researcher to produce quantitative statistics about certain populations (Fowler, 2014), they are appropriate for collecting information about social trends, consumer opinions, or consumer attitudes (Malhotra and Birks, 2007; Bryman and Bell, 2011).
When constructing a survey, there are both pros and cons to consider. Looking into the pros, Malhotra and Birks (2007) and Bryman and Bell (2011) argue that a questionnaire is simple to administer to a large sample, and that the data collected is consistent because of the structured set-up of the questions which limits the variability in the results. Fowler (2014) argues that when creating a self-administered questionnaire, it is important that the researcher uses closed question that makes it easy for the respondents to answer by clicking in one box and does not leave the respondents presuming. Further, the coding, analysis and interpretation of data are simple. However, when administering a questionnaire, the researcher needs to consider cons such as misunderstandings by the respondents, the lack of elaborate answers, respondent fatigue, missing data and low response rates (Malhotra and Birks, 2007; Bryman and Bell, 2011).
explaining the nature of the study, contact details to the researcher and an assurance of the respondents’ anonymity.
The survey has several questions regarding each factor identified from the literature review. All the questions presented in the questionnaire were predefined and taken from relevant theories as described in the operationalization that can be found in table 2. All the questions were set to be mandatory to answer in order to continue. The main questions regarding the different factors were measured on a Likert scale from "1 Strongly disagree" to "7 Strongly agree". The last section of the questionnaire contained the control questions. The full questionnaire can be found in Appendix II.
For the question design, the respondents were asked about their perception of each factor, meaning that their answers relate to their expectations of PSS. Grönroos (2007) argues that expectations are important and plays a crucial role in the perceived quality of the service, which relates the satisfaction of the consumer. Further, it is important to know that these expectations are made up before the actual experience (Olson and Dover, 1979). Therefore, one can arguably say that these factors are important for the success of PSS.
3.4.1 Pre-study
In order to know within what field of consumer goods this study was to focus on, a pre-study was conducted. On the 15th of February 2018, a focus group was conducted with
seven participants from a variety of backgrounds. The focus group started with a short explanation of the concept of CBM an PSS. After that, the researcher introduced some different options that the participants could discuss freely. The conversation was self-sufficient, and the researcher only had to ask a few questions. The group agreed unanimously on home electronics as the topic they would be most interested in for trying PSS. The questions and some quotes can be found in Appendix II.
3.5 Data Collection Instrument
In this chapter, the operationalization that was the basis for the development of the questionnaire is presented.
3.5.1 Operationalization
Table 2 - Operationalization of Theoretical Concepts
Concept Theoretical definition Questions Item
Economical Value
Edbring et al. (2016) argue that the economic reasons are among the main motivators for PSS. Cost does not merely relate to the monetary price the consumer has to pay cost also entails behavior (Tobler et al. 2012). Further, there are several factors that has an impact on our cost behavior when it comes to environmental benefit (Tobler et al. 2012). These can be time, discomfort, effort and other subjectively defined inhibitions of one’s behavior (Diekmann and Preisendörfer, 1998; Kollmuss and Agyeman, 2002; Tobler et al. 2012).
1. I perceive that PSS
would be cheaper over time
2. I perceive that PSS
arrangements would provide more value for me
3. I perceive that PSS
would be beneficial for me
4. I perceive that PSS
would be time efficient
5. I perceive that PSS
would be comfortable
6. I perceive that PSS
would be effortless
Seven-point Likert scale
1 = Strongly disagree 7 = Strongly agree
Measurement 1. Price - Edbring et al.
(2016), Liu et al. (2017)
2. Value (Liu et al. (2017) 3. Product benefit - Barber
et al. (2016)
4. 5. 6. Time, Discomfort,
Effortless - Diekmann and Preisendörfer (1998), Kollmuss and Agyeman (2002), Tobler et al. (2012)
Flexibility
Flexibility is highly related to the satisfaction of the consumer in a PSS (Rexefelt and Hjort af Ornäs, 2009; Bardhi and Eckhardt, 2012; Edbring et al. 2016). It provides a sense of freedom to be able to use and utilize the products rather than owning them (Catulli et al. 2013). PSS provides the flexibility that the ridged institution of ownership lack (Edbring et al, 2016). The opportunity and variability are a strong motivator in products that require maintenance and upgrade, for example products with fast innovation cycles (Rexefelt and Hjort af Ornäs 2009; Edbring et al. 2016).
7. I perceive that PSS
could give me the freedom to choose the products I need
8. I perceive that PSS
would not require a big commitment 9. I perceive that PSS would increase my opportunity to switch products as often as I need 10. I perceive that PSS
could provide access to the products I need
11. I perceive that PSS
could give me the freedom to choose between different levels of product and services
Seven-point Likert scale
1 = Strongly disagree 7 = Strongly agree
Measurement
7. Freedom - (Catulli et al.
(2013), Edbring et al. (2016)
8. Commitment - Rexefelt
and Hjort af Ornäs (2009)
9. Variability - Rexefelt and
Hjort af Ornäs (2009), Edbring et al. (2016)
10. Guaranteed access -
Edbring et al. (2016)
11. Flexibility - Rexefelt
and Hjort af Ornäs (2009), Edbring et al. (2016)
Trust
In PSS, there is a relationship between the consumer and the provider, which comes with some attachments. To form positive attitudes towards a PSS provider, the consumer needs to trust the provider (Catulli et al. 2013; and Armstrong et al. 2015). Since PSS is a rather new way to consume, there is a lot of perceived risk in the consumer’s mind (Edbring et al. 2016). However, Baumeister (2014) argues that this perception decreases as the consumer gets more accustomed to PSS and start to trust the relationship.
12. I perceive that PSS
companies would be able to create a customer relationship with me
13. I perceive that I would
be able to trust PSS arrangements
14. I perceive that PSS
arrangements are honest
15. I perceive that PSS
arrangements would provide satisfactory information
Seven-point Likert scale
1 = Strongly disagree 7 = Strongly agree Measurement 12. Relationship - Baumeister (2014) Raja et al. (2013), Edbring et al. 2016
13. Trust - Edbring et al. (2016) 14. Openness - Raja et al.
(2013), Edbring et al. (2016)
15. Information - Luhmann
(1979)
Desire to Own
Product ownership is one of the main characteristics of the modern consumption culture (Wallendorf and Arnould, 1988). This means that there is an institutionalized social norm which has taught generations of the importance of ownership. Catulli (2012) and Edbring et al (2016) both concluded that ownership is an inherent value in the modern consumption society and that it has an important impact on how people view
16. It is important for me
to own my products
17. Products I own reflect
who I am
18. I perceive that PSS
products would require careful treatment compare to the products I own
19. I perceive that PSS
products would include some sanitary concerns
Seven-point Likert-scale
1 = Strongly disagree 7 = Strongly agree
Measurement
16. Importance of ownership -
Wallendorf and Arnould (1988)
17. Self-image - Solomon et al.
(2007)
18. Handle with care - Edbring
et al. (2016)
Peer Influence
Consumer socialization is a process by which the consumers acquire knowledge and attitude towards their role in the consumer market (Ward, 1974). A part of this learning process is the social interactions and the knowledge gathered from peers (Bandura, 1969). This peer influence is recognized as one of the main factors influencing consumer behavior (Churchill and Moschis, 1979;
Mangleburg et al. 2004; Luo, 2005; Baek and Ho, 2015).
20. My friends’ purchases
influence my own purchases
21. I can justify my
purchases based on what other people around me do
22. The products I own,
or lease reflects my place in the society
23. I can express my
belonging to a group through the products I own
Seven-point Likert scale
1 = Strongly disagree 7 = Strongly agree
Measurement
20. Guidance - Churchill and
Moschis, (1979), Mangleburg et al. (2004), Luo, (2005), Baek and Ho (2015)
21. Justification - Beak and Ho
(2015)
22. Social Identity - Hirschman
and Holbrook, (1982),Beak and Ho, (2015)
23. Expressive needs -
Mangleburg et al. (2004), Beak and Ho (2015)
Willingness to Adopt
When discussing the willingness to adopt, one needs to consider that there are several different stages of adoption. In the knowledge phase, the consumers get exposed to the existence of the product or service and gain understanding about it (Rogers, 2003; Sahin, 2006). However, it is in the next stage, the persuasion phase, that the willingness of the adoption process is tested (Rogers, 2003; Sahin, 2006). Sahin (2006) argues that it is during the persuasion stage that the consumers weigh the perceived characteristics or attributes of the innovation, which affect the willingness.
24. There are advantages
in trying PSS
25. PSS would be
compatible with my life
26. PSS would be easy to
implement in my life
27. I would like the
opportunity to try PSS
28. Using PSS would
have a visible impact in my life Seven-point Likert-scale 1 = Strongly disagree 7 = Strongly agree Measurement 24. Advantage - Rogers (1995), Rogers (2003) 25. Compatibility - Rogers (1995), Rogers (2003) 26. Complexity - (Rogers (1995), Rogers (2003) 27. Trialability - Rogers (1995), Rogers (2003) 28. Observability - Rogers (1995), Rogers (2003) Environmental Awareness
In previous literature, there has been some contradicting studies regarding how environmental awareness and the consumers’ attitudes towards the environment have an impact on the consumption. Liu et al. (2017) argue that there are several different layers to this factor.
A consumer with positive environmental attitudes would be willing to spend more money to uphold a better environmental behavior (Birgelen et al. 2009; Dono et al. 2010; Gadenne et al. 2011). Bartiaux (2008) argues that there is no connection between environmental knowledge and action. Previous literature argue that there is a gap between a consumer’s environmental awareness and the actual behavior.
29. I believe that our
society is going towards a more sustainable future
30. I believe that we need
to change our way of consuming in order to preserve the environment
31. I am aware of the
environmental challenges that our society is facing and I want to contribute to a sustainable future Seven-point Likert-scale 1 = Strongly disagree 7 = Strongly agree Measurement 29. Pro-environmental beliefs -
Gadenne et al. (2011), Liu et al. (2017)
30. Willingness to pay -
Birgelen et al. (2009), Dono et al. (2010), Gadenne et al. (2011)
31. Environmental
consciousness - Arslan et al. (2012), Assarut and Srisuphaolarn (2012)
Diffusion of Innovation
There are three main categories, early
adopters, majority (early/late), and laggards (Holak et al. 1987; Armstrong et
al. 2012). Early adopters are among the first to try a new product and are in some cases seen as an opinion leader (Solomon et al. 2013). After the early adopters, the majority adopts the product, this group of people can sometime wait deliberately
32. What type of consumer are you? - I like to stay updated
and buy the latest products as soon as possible
- I wait for a while before I buy new products, depending on price and
Nominal scale Measurement 32. - 1. Early adopters 2. Majority 3. Laggards
3.5.2 Pretesting
After the operationalization, the questionnaire was first sent out to different university professors within the field of study to ensure the quality. Then, a pre-test was performed in the target population – approximately 15 responses were collected with different comments for improvement or clarification. Overall, the feedback was positive. However two questions were added and a few were modified to make sure that the participants understood the questions. The questionnaire was then distributed on Facebook and on a platform for research.
3.6 Data Analysis Method
In this study, several different data analysis methods were used in order to interpret the data. It is important to know early in the research process what type of data analysis is needed to fulfill the purpose of the research (Bryman and Bell, 2011; Fowler, 2014). For this study, several different methods were applicable. Before the analysis took place, the answers collected in the questionnaire were coded and organized in order to quantify and analyze them. This was done by having the numbers matched with specific answers (Malhotra, 2010; Bryman and Bell, 2011).
All the data was analyzed in the software program IBM SPSS Statistics (SPSS). SPSS supports the entire analytical process, including instruments useful for arranging the data collection and organizing the results in an understandable way (IBM, 2017). The first data analysis for this research was a descriptive analysis, including frequency, mean, and standard deviation of the variables in order to get an understanding of the sample population (Bryman and Bell, 2011).
3.6.1 Quality Criteria
When evaluating a study, to reduce measurement errors, the researcher needs to address two important characteristics: validity and reliability (Hair et al. 2010). Validity refers to the degree of which a measure accurately represents what it is supposed to represent (Hair et al. 2010). When validity is assured, the researcher also needs to consider the reliability which is how “true” or “error free” the measurement is, meaning that it estimates the consistency (Hair et al. 2010; Bryman and Bell, 2011).
As argued above, validity concerns whether the components measure what they are supposed to measure (Bryman and Bell, 201; Hair et al. 2015). Furthermore, the content, criterion and construct validity are assessed in order to assure the validity of the research (Malhotra, 2010). Some argue that construct validity is the most difficult to establish since it relates directly to the appropriateness of the component (Malhotra, 2010; Bryman and Bell, 2011). To establish construct validity and external validity, the researcher needs to build a comprehensive literature review and a thorough description of the construct used as well as gain understanding of what interpretations can be made (Malhotra, 2010). To ensure this, a detailed and sound evaluation of previous literature was used to formulate the operationalization. Further, a close discussion with professors was done to avoid discrepancies. After that, the content validity, which examines whether a measurement accurately represents the concept (Bryman and Bell, 2011; Malhotra 2010), was ensured through a pre-test and further discussions with professors within the field (Hair et al. 2010). Finally, a correlation test (Pearson’s r) was used to strengthen the criterion validity, which refers to the fact that the constructs might measure the same thing (Bryman and Bell 2011).
Reliability, which concerns the stability of a measurement, can be both external and internal. Bryman and Bell (2011) and Golafshani (2003) argue that reliability establish that the study is stable over time. It evaluates the measurements’ consistency and the result’s repeatability (Bryman and Bell, 2011; Malhotra, 2010). To achieve external reliability, a study needs to be conducted in an unbiased and impartial manner – any explanations should be objective and based on facts. Furthermore, the measurements should be consistent to give accurate results (Bryman and Bell, 2011; Malhotra, 2010). Bryman and Bell (2011) and Golafshani (2003) also argue that a well-chosen sample of the population increases the reliability of the study and its stability.
To achieve internal reliability, the scales used in the study were tested for consistency. This means that the answers that the respondents give in different questions about the same concept should be related. The reliability coefficient, which assesses the consistency of the entire scale, can statically be established using Cronbach’s Alpha (Hair et al. 2010; Malhotra, 2010). In this research, Cronbach’s Alpha is tested for each measurement, which means that the concept is stable and that it measures the same thing (Hari et al. 2010). The limit accepted for the study was set to 0.6 (Malhotra and Birks, 2006).
3.6.2 Hypothesis Testing
The R square (R2) is examined which is illustrated by a value between -1 and +1, where 0 represents no relationship between the variables (Malhotra and Birks, 2006). In the hypothesis testing the adjusted R2 is used to determine the relationship between the
dependent and independent variables since it includes the number of independent variables and the sample size in the equation. Further the beta value was examined to know how much the independent variable interferes with the dependent variable (Nardi, 2003). In the hypothesis testing including the moderator, the aim is to investigate if the interaction of a third variable (X2) on the independent variable (X1) changes the relation towards the dependent variable (Y). Hancock and Mueller (2010) explain that the relation between X1 and Y might depend on X2 in which case X1 and X2 are to interact. They further argue that the ability to modify a model is critical since many theories within social and behavioral science require that the relation between two values depends on a third. In this study, a significance level is set to at least 95% that is represented by the value of .05 for the hypothesis testing, this is recommended in marketing research (Nardi, 2003).
In order to provide additional findings to this study, an Analysis of Variance (ANOVA) was done to compare the mean of the groups. One of the control questions relates to the diffusion of innovation and the respondent’s self-evaluation of their adoption rate, thus dividing them into three different groups. According to Hancock and Mueller, (2010) the ANOVA is applicable when the research is done to determine if an independent variable causes variation in the outcome of the dependent variable. The independent variable, or respectively the categorical, nonmetric variable, will be compared to the dependent variable. In this study, this is done by an interval measurement scale (Malhotra and Birks, 2006). The comparison of the mean values of the dependent variable was done through the one-way ANOVA. This included Degrees of Freedom (df), Sum of Squares, Mean Square F and Sig-level.
3.7 Ethical Considerations
When conducting a study, it is important to consider the ethical issues regarding data collection from participants. Bryman and Bell (2011) argue that there are some main concerns that need to be considered. First of all, no harm should come to the participants or the researcher, this can be physical or psychological harm (Nardi, 2003; Bryman and Bell, 2011). It is the researchers’ task to ensure a safe environment and reduce any risk (Merriam, 2009; Bryman and Bell, 2011). Further, the researcher should not in any way deceive the participants, and a clear explanation of the expectations should be presented to the respondents (Bryman and Bell, 2011). This can concern the nature, length or purpose of the study.
participants’ privacy, including ensuring anonymity and confidentiality of the participants and respondents (Bryman and Bell, 2011).
4 Data Analysis
In this chapter, the result of the data testing is presented and analyzed. It starts with an analysis of the sample, followed by the factor analysis to ensure validity and control the components. Then the reliability testing is presented, followed by another validity test. Finally, the hypothesis testing is presented. All of the tables can be found in Appendix III.
4.1 Descriptive Statistics
In total, the study has 463 respondents. After going over the data, none of the respondent’s data was taken out, keeping all of them for further analysis. The sample was drawn from the population which did not have many restrictions. However, since both Facebook and a platform for research was used, the sample was set to be collected from Europe. The largest age group that answered the questionnaire was between the ages of 25-34 with 37.4%, followed by the ages 35-45 with 21.2%. The vast majority, 67.8%, of the respondents were employed full-time and when it comes to gender, 50.3% were female and 47.7% were male.
Further, the respondents were asked to consider what type of consumer they were in terms of diffusion of innovation. Most of them considered themselves as the majority, 51.2%, followed by laggards, 35.2%, and lastly the early adopters, 13.6%.
All the descriptive sample information can be found in appendix III in tables 5 to 8.
4.2 Validity and Reliability Analysis
For the validity testing, the data was analyzed with an exploratory factor analysis (EFA), meaning that it was analyzed by a principal component analysis. This analysis gave new information and insight to the data, and there were some factors that needed consideration before moving on. As one can se in table 4 below, Flexibility and Trust loaded on each other as one component, even though the literature review clearly argues that they should be treated separately. To further test this, a confirmatory factor analysis (CFA) was conducted, which included only the items from flexibility and trust. There was one item that did not load on any other, and it was therefore taken out before this test. This resulted in a clear division of the components and that another item was excluded from further analysis. To fully ensure that these items and components were separate, a correlation analysis was done using the Pearson’s r test. It showed that none of the items from the different components had a high correlation coefficient, meaning that there is no multicollinearity between them.
Another finding of the EFA, was that the component of peer influence actually consists of two separate components. Therefore, peer influence was divided into two separate factors, each containing two items. The first component includes guidance and
justification and the second one includes social identity and expressive needs. Further,
including the CFA and the correlation test can be found on Appendix III, tables 10 and 11.
Table 3 - Premliminary Rotated Component Matrix
Since the EFA showed that Peer Influence actually consists of two factors/independent variables, the data analysis from this point changed from five to six factors that has an impact on the willingness to adopt PSS. When looking closer into the separation of the items, there was a clear division between them. The first factor relates to the decision making and the second relates to the social belonging and expressiveness. Therefore, a new model was constructed with a new set of hypotheses related to peer influence. All the other factors remained the same.
H5a: Peer Related Decisions of PSS has an impact on the consumers’
willingness to adopt PSS
H5b: Social Influence of PSS has an impact on the consumers’ willingness
to adopt PSS Component 1 2 3 4 5 FLEX_guarant_access .809 FLEX_flexibility .768 FLEX_freedom .742 FLEX_variability .732 TRU_infomration .686 TRU_trust .659 TRU_openess .634 TRU_relationship .612 FLEX_commitment ECV_value .855 ECV_prive .845 ECV_product_benefit .838 ECV_time .715 ECV_discomfort .676 ECV_effortless .558 PEER_guidance .902 PEER_justification .853 DTO_selfe_ima .853 PEER_social_identidy .718 PEER_expressive .519 .708 DTO_sanitation .815 DTO_handle_care .738 DTO_importance_t_own .585
Extraction Method: principal Component Analysis Rotation Method: Varimax with Kaiser Normalization KMO: .907
Figure 2 - New Conceptual Model After the Factor Analysis
The reliability testing resulted in the exclusion of one item of the environmental awareness. Otherwise, as displayed in table 4 below, all factors had a Cronbach’s Alpha above .6 which was the set limit for this study. Economical value had the highest value at .914 whereas the desire to own had the lowest at .612.
Table 4 - Reliability Testing
ECV FLEX TRUST DTO PEER 1 PEER 2 ADO ENV
Cronbach’s
Alpha 0.914 0.851 0.880 0.612 0.873 0.882 0.892 0.810
Items 6 4 4 3 2 2 5 2
EVC – Economical Value, FLEX – Flexibility, TRU – Trust, DTO – Desire to Own, PEER1 – – Peer Related Decisions, PEER2 – Social Influence, ADO – Willingness to Adopt, ENV – Environmental Awareness
Finally, the validity was tested for the different factors using a Persons’ r correlation test. None of the independent variables showed a high correlation coefficient, which means that there is no multicollinearity between the different independent variables. The result of the test is shown in Appendix III, table 14.
4.3 Hypothesis Testing
accepted, the Beta value needs to be either positive or negative depending on the nature of the hypothesis. It also needs to be statistically significant.
Table 5 - Regression Analysis Without Moderation
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8
Beta Beta Beta Beta Beta Beta Beta Beta
Control variables
Gender -.019 -.018 -.077 -.037 -.036 .030 .002 -.38
Age -.133 -.008 -.158** -.099** -.024** .010 -.039 -.027
Occupation -.023 -.072 -.032 .014 -.024 -.014 -.041 -.045
Determinants of Willingness to Adopt PSS
H1 – Economical Value .825** .565** H2 – Flexibility .602** .158** H3 – Trust .711** .206** H4 – Desire to own -.163** -.087** H5a – Peer 1 .367** .083* H5b – Peer 2 .281** -.012 R2 .021 .688 .379 .523 .047 .130 .091 .753 Adjusted R2 .015 .686 .374 .519 .039 .123 .083 .748 Std. error of the estimate 1.35 .762 1.077 .943 1.334 1.274 1.302 .682 F-Value 3.332 253.063 69.884 125.740 10.063 27.855 11.517 153.617 DF. 3 4 4 4 4 4 4 9 ** sig at 0.01 level *sig at 0.05 level
In Appendix III in tables 15 to 20, the regression analysis for the moderated variables is presented. Model 1 only contains the control variables, followed by the independent testing of the moderated factor. Finally, model 3 is the full model testing where it is determined if the factor is either accepted or rejected. In table 4, a summary of the hypothesis is presented. It displays both when each independent variable is tested independently towards the willingness to adopt PSS and when it is tested in a full model multiple regression. Seven out of the twelve hypotheses were accepted. For the first set of hypotheses, without the moderator, only social influence got rejected since it was not statistically significant. For the hypothesis including the moderation, only HP10a and HP10b were accepted.
Table 6 - Summary of the Regression Analysis
Tested independent Full Model Testing
HP Beta Adjusted R2 Beta Adjusted R2 Accepted/Rejected
H1 . 825** .686 .565** .748 Accepted
H2 .602** .374 .158** .748 Accepted
H3 .711** .519 .206** .748 Accepted
H4 -.163** .039 -.087** .748 Accepted
