Access-Based Consumption: A Choice-Based Conjoint Analysis for
Determining Consumer Motives for (not) Engaging in Car Sharing.
Access-Based Consumption: A Choice-Based Conjoint Analysis for
Determining Consumer Motives for (not) Engaging in Car Sharing.
by Erin Hiemstra January 16, 2017 Master Thesis MSc Marketing, Intelligence University of Groningen Faculty of Economics and Business
Department of Marketing PO Box 800, 9700 AV Groningen
Supervisors:
First Supervisor: dr. F. Eggers Second Supervisor: prof. dr. P.C. Verhoef
Author: Name: Erin Hiemstra Student number: S2199580
Address: Madoerastraat 5b, 9715HE Groningen Phone number: +31612473547
Management Summary
Nowadays a shift from ownership to sharing takes place. Research on the consumer preferences of access-based services is limited. Lots of traditional theories need to be investigated in this context. Car sharing is used in the study as an application of access-based consumption. In the Netherlands consumers are barely making use of car sharing. Therefore, the outcomes may also be interesting for car sharing firms. This study will investigate to what extent car sharing attributes have an impact on consumers their (un-)willingness to engage in car sharing. The impact of attributes like product brand, price per hour, drive train and service characteristics is not heavily investigated in this context. Furthermore, the research will focus on whether differences between consumers their (un-)willingness to engage in car sharing could be explained by contagion, need for status, consumer materialism and environmental concern.
The study will apply an aggregate Logit and a hierarchical Bayes (HB) choice-based conjoint model in order to investigate these issues. HB is used to account for individual heterogeneity. The findings indicate that consumers prefer a low price per hour the most. Furthermore, with regard to the service characteristics, consumers think that flexibility is important. They prefer picking-up and returning the car anywhere within the city above returning the car to a fixed station. Reservation is also an important attribute to consumers. Consumers tend to favour making no reservation above making a short- or long-term reservation. Besides, prestige car brands are more liked than volume or price car brands. Also the electric and hybrid vehicles are preferred rather than conventional fuel brands. Furthermore, pro-environmentally oriented consumers are most likely to engage in car sharing. Contagion, consumer materialism and need for status do not have a significant impact on the willingness to engage in car sharing.
sharing firms are recommended to make use of green vehicles. If a car sharing company makes use of green vehicles, consumers are more likely to engage in car sharing. Pro-environmentally oriented consumers, consumers that are involved with cars, males and consumers that have a high amount of car sharing knowledge are most likely to engage in car sharing. Therefore, managers might focus on these consumers.
This study is subject to limitations that leave avenues for further research. Firstly, a conjoint analysis has several drawbacks. Therefore, the results might be different from real market experiences. Secondly, it only uses car sharing as an application for access-based services. This reduces the possibility to generalize the results to access-based services. Thirdly, the data sample (n=208) is relatively small and consists of many young and well-educated consumers. Fourthly, it uses only one car brand per brand category. This makes it difficult to generalize the results to the car brand categories. Fifthly, it only evaluates price per hour and ignores other components of the cost structure. Sixthly, it does not focus on car sharing firms as a brand. Seventhly, the study does not make use of an existing scale of consumer-to-product contagion. Measures of contagion are not established yet. Finally, it does not use a scale of possessiveness in order to investigate the impact of materialism on engaging in car sharing.
Preface
The thesis is written to finalize my master in Marketing Intelligence at the University of Groningen. In September 2015 I started with my master Marketing after first completing the bachelor Business Administration. During my bachelor I developed a passion for marketing related topics. After doubting about the track within the master Marketing, I decided to continue with Marketing Intelligence instead of Marketing Management. Writing a thesis is always challenging, since it is the first time that you conduct a research yourself. Therefore, the process of writing the thesis had some ups and downs. Despite the fact that the research period was intense, I really have the idea that I learned a lot and that I was able to develop my research skills. I would not have been able to write this thesis, without the help of several people. First of all, I would like to thank dr. Felix Eggers for his high-quality feedback and great help during the process of writing this thesis. He was always available to help me and answer questions. Secondly, I want to thank all the people that helped me collecting the data. In particular dr. Felix Eggers, who made it possible to distribute the survey among students of the course Marketing Engineering. Thirdly, I want to thank my friends for the great time I experienced in Groningen. Finally, I want to say thanks to my family who are always really supportive in everything I do. This thesis is not only the end of my master program, but also the end of being a student. I have experienced great years at the University of Groningen. During these years I have learned a lot. I look forward to bring this in practice in the upcoming period.
Table of Content
MANAGEMENT SUMMARY ... 3
PREFACE ... 5
1. INTRODUCTION ... 8
2. THEORETICAL FRAMEWORK ... 10
2.1 ACCESS-‐BASED CONSUMPTION ... 11
2.2 CAR SHARING: AN APPLICATION OF ACCESS-‐BASED CONSUMPTION. ... 12
2.3 HYPOTHESES ... 15
2.3.1 Car Brand ... 16
2.3.2 Price per Hour ... 16
2.3.3 Drive Train ... 17
2.3.4 Reservation ... 18
2.3.5 Vehicle Station System ... 18
2.3.6 Moderators ... 19
2.3.6.1 Contagion ... 19
2.3.6.2 Environmental Concern ... 20
2.3.6.3 Consumer Materialism ... 21
2.3.6.4 Need for Status ... 22
2.3.7 Control Variables ... 23
3. METHODOLOGY ... 24
3.1 METHOD OF PREFERENCE MEASUREMENT ... 25
3.2 ESTIMATION PROCEDURE ... 25
3.3 STUDY DESIGN ... 25
3.3.1 Attributes and Levels ... 25
3.3.2 Measures ... 27
3.3.2.1 Contagion ... 27
3.3.2.2 Environmental Concern ... 27
3.3.2.3 Consumer Materialism ... 27
3.3.2.4 Need for Status ... 28
3.4 CHOICE DESIGN ... 28
3.5 THE MODEL ... 29
4. EMPIRICAL STUDY ... 30
4.1 CHOICE ELICITATION: SURVEY ... 30
4.1.1 Data Collection ... 30
4.1.2 Procedure ... 31
4.2 SAMPLE CHARACTERISTICS ... 31
4.3 DIMENSION REDUCTION ... 33
4.4 AGGREGATE CHOICE-‐BASED CONJOINT ANALYSIS ... 33
4.4.1 A Part-‐worth or a Linear Model? ... 34
4.4.2 Model Fit ... 34
4.4.3 Main Effects ... 35
4.4.4 Moderating Effects ... 37
4.4.5 Predictive Validity ... 39
4.5 HIERARCHICAL BAYES CHOICE-‐BASED CONJOINT ANALYSIS ... 39
4.5.1 Model Fit ... 40
4.5.2 Main Effects ... 40
4.5.3 Moderating Effects ... 42
4.5.4 Control Variables ... 43
4.6 HYPOTHESES TESTING SUMMARY ... 44
5. DISCUSSION ... 45
5.1 CAR SHARING PREFERENCES ... 45
6. SUMMARY ... 48
6.1 THEORETICAL IMPLICATIONS ... 49
6.2 MANAGERIAL IMPLICATIONS ... 50
7. LIMITATIONS AND FURTHER RESEARCH ... 51
REFERENCES ... 53
APPENDIX A – CAR SHARING PRICES ... 66
APPENDIX B – MEASURES CONTROL VARIABLES ... 66
APPENDIX C – EXEMPLARY CHOICE SET ... 68
APPENDIX D – HOLDOUT CHOICE SET ... 68
APPENDIX E – TABLES PRINCIPAL COMPONENT ANALYSIS AND RELIABILITY ANALYSIS .. 69
APPENDIX F – UTILITIES AND P-‐VALUES MODEL 1, 2 AND 3 ... 69
1. Introduction
Ownership and possessions have been popular topics since the 50s in the marketing literature. Snare (1972) indicates that ownership is expressed by an extraordinary long-term relationship between an individual and an object. Temporal or long-term use of a product without purchasing and owning it is referred to as access or sharing in the current stream of literature (Chen 2009; Bardhi & Eckhardt 2012). In the past owning an object was seen as a superior consumption mode (Ronald 2008). Sharing a product with others was perceived as a waste of one’s money, not making use of someone’s purchasing power and limiting the freedom. Owning a product had a positive impact on people’s self-concept and represented a symbol of being autonomous (Richins 1994; Kleine, Kleine & Allen 1995). Consumers perceived an object as a part of the extended self (Belk 1988). The status of consumers who were sharing a product was perceived as being lower, since they were primarily financially motivated (Durgee & O’Connor 1995).
Nowadays sharing is getting a more prominent role in the lives of consumers. Therefore, it is becoming an important topic within the marketing domain (Belk 2007; Belk 2010). Ownership is no longer an ultimate expression of consumer desire, as a result of which a shift from ownership to sharing takes place (Chen 2009). Research on product sharing systems bears interest to a bundle of related consumption practices, such as: access-based consumption (Bardhi & Eckhardt 2012; Schaefers, Lawson & Kukar-Kinney 2016; Lawson et al. 2016), collaborative sharing (Belk 2007), commercial sharing systems (Lamberton & Rose 2012) and collaborative consumption (Botsman & Rogers 2010). Access-based consumption is the most popular term for product sharing systems. This consumption mode is also present in the mobility sector (Bardhi & Eckhardt 2012). In this sector the focus shifts from privately owned cars to the use of mobility services (Nykvist & Whitmarsh 2008). Car sharing is such a mobility service.
lead to uncertainty. Therefore, the desirability of car sharing may be reduced (Hildebrandt et al. 2015). Despite the fact that ownership becomes less important, the intention to use and diffusion rate of car sharing in Europe remain rather low (Shaheen & Cohen 2013; Ornetzeder & Rohracher 2013). In the Netherlands consumers are barely making use of such mobility concepts (Moeskops 2015). This study will primarily focus on the Dutch business-to-consumer market.
be important since these services are associated with less exclusiveness (Lamberton & Rose 2012). Therefore, status-oriented consumers might be less likely to engage in car sharing. Besides, Bardhi and Eckhardt (2012) argue that contagion might be an important motive for not engaging in car sharing. Argo, Dahl and Morales (2006) state that contagion is a form of disgust in which consumers are aware that other consumers physically touched an object. Recent studies indicate that pro-environmentally oriented consumers are willing to engage in car sharing (Lawson et al. 2016). However, Bardhi and Eckhardt (2012) do not find evidence for this relation. This research will study the impact of contagion, environmental concern, consumer materialism and need for status on the willingness to engage in car sharing.
This study contributes to the literature in several ways. Firstly, little research has been conducted to access-based services. Existing theories on consumer behavior need to be investigated in this context (Chen 2009). Secondly, it helps understanding the drivers of engaging in car sharing. Thirdly, it takes a branding and pricing perspective. Lawson et al. (2016) indicate that the impact of price needs to be investigated in this context. The impact of product brand has not been studied extensively. Peck and Shu (2009) indicate that it still might be important, whereas others propose that the importance is reduced significantly (Bardhi & Eckhardt 2012). Fourthly, it takes into account the drive train and service attributes. Fifthly, it accounts for heterogeneity in consumer preferences by using a hierarchical Bayes conjoint analysis. Finally, it determines possible causes why consumers do (not) engage in car sharing. The study tries to answer the following two research questions in order to determine the preferable attributes of car sharing and to establish motives of consumers for (not) engaging in car sharing:
“To what extent do car sharing attributes have an impact on consumers their (un-)willingness to engage in car sharing?”
“To what extent could differences with regard to consumers their (un-)willingness to engage in car sharing be explained by contagion, environmental concern, consumer materialism and need for status?”
The remainder of the paper is structured as follows. Firstly, this study will review literature dealing with access-based consumption and car sharing as well as with developing a conceptual framework based on existing theories. Secondly, the methodology of the conjoint analysis is presented. Furthermore, also the modeling approach and the way the data is collected will be discussed. Thirdly, the results are analyzed and presented. Finally, the theoretical and managerial implications, limitations and possible directions for further research are described.
2. Theoretical Framework
be presented. Moderators like contagion, environmental concern, consumer materialism and need for status will be used to develop theories why consumers do (not) engage in car sharing. Finally, the control variables and the conceptual model of this study will be discussed.
2.1 Access-based Consumption
An important trend in research on consumer behavior is that ownership is not an ultimate expression of desire anymore (Chen 2009). Several concepts have been introduced for product sharing systems. The most popular concept is access-based consumption. Bardhi and Eckhardt (2012, p. 881) define it as follows: “access-based consumption is about transactions that may be market mediated in which no transfer of ownership takes place.” According to this definition, consumers acquire an object temporarily and pay a price for using the object. The owner is the individual or firm that offers the service. This definition is used in several recent studies (Belk 2014; Schaefers, Lawson & Kukar-Kinney 2016; Lawson et al. 2016; Schaefers et al. 2016). Access-based services allow consumers to avoid the ‘burdens of ownership’. The risks of ownership (financial, performance and social risk) enhance access-based consumption usage (Schaefers, Lawson & Kukar-Kinney 2016). Therefore, consumers are not primarily focused on purchasing and possessing objects anymore. They want access to products and are willing to pay for this experience (Bardhi & Eckhardt 2012). Lawson et al. (2016) indicate that consumers might engage in access-based consumption based on economic and environmental grounds. Furthermore, they indicate that status may be a motive to engage in access-based services. Whereas Chen (2009) expected that a reduction in status-orientation would lead to more interest in access-based services. According to Botsman and Rogers (2010) access-based services may be beneficial for the environment. Products that previously have been used infrequently and were privately owned are shared among multiple consumers to maximize the usage. The rise of access-based consumption modes might be explained by the theory of perceived ownership (Belk 2010). Perceived ownership is referred to as a feeling that something is “mine”, without owning the object (Pierce, Kostova & Dirks 2001; Peck & Shu 2009). Consumers may incorporate objects in their extended self and feel a sense of perceived ownership by touching objects, controlling or creating them or by becoming familiar with an object (Belk 1988; Pierce, Kostova & Dirks 2001; Peck & Shu 2009). However, some consumers are still more attached to owning their objects (Bardhi & Eckhardt 2012; Lawson et al. 2016).
Baumeister, Scherer and Van Wangenheim (2015) studied the importance of product brands in sharing systems. They found that the importance of product brands might be reduced for some access-based services. However, they did not investigate how important this attribute is compared to the other attributes. Therefore, this study will investigate how important product brands are. Besides, the impact of price is not heavily investigated in the context of access-based services (Lamberton & Rose 2012; Lawson et al. 2016). This research will study the importance and impact of price in this context. Several studies indicate that service characteristics of sharing systems might be important for consumers to decide whether they will make use of access-based consumption or not (Lamberton & Rose 2012; Baumeister, Scherer & Van Wangenheim 2015). Nonetheless, these studies do not indicate which attributes are important for consumers in the access-based consumption applications. This study will focus on these aspects.
2.2 Car Sharing: an application of access-based consumption.
The shift in focus from ownership to sharing supports the use of car sharing programs. In the past consumers were strongly attached to their cars (Ball & Tasaki 1992). This symbolized their identity and was associated with masculinity and independence. Nowadays within the mobility industry the focus shifts from privately owned cars to the use of mobility services to fulfil individual mobility demands (Nykvist & Whitmarsh 2008; Bardhi & Eckhardt 2012).
Bardhi and Eckhardt (2012, p. 886) define car sharing as “a group of paying individuals who gain access to a fleet of cars along with other paying members periodically over time”. Car sharing models vary primarily along two dimensions: the vehicle station system and the necessity to make a reservation (Boyacı, Zografos & Geroliminis 2015; Kaspi, Raviv & Tzur 2014; Firnkorn & Shaheen 2016). Some car sharing models make use of a fixed car sharing station in which consumers have to pick-up and return the vehicles. Other concepts offer the consumers the opportunity to pick-up the car at a specific station and drop-off the car at another car sharing station. Finally, there are also concepts in which the car sharing company allows the user to pick-up and drop-off the car at any location within the city where the firm is active. These firms make use of GPS to locate the cars. Furthermore, there are differences with regard to reservations. Some car sharing concepts require a short- or long-term reservation, whereas this is not necessary for other car sharing models.
2010). This study will focus on car brand, price per hour, drive train, reservation and vehicle station system. In the section below the theoretical relevance of the attributes is explained.
In the past consumers were able to maintain stable relationships with brands and expressed their identity by purchasing certain brands (Muniz & O’Guinn 2001; Schouten & McAlexander 1995; Moore 2004). The rise of access-based consumption might alter the importance of brands and increase the importance of services according to several studies (Durgee & O’Connor 1995; Lovelock & Gummesson 2004; Bardhi & Eckhardt 2012). These researchers proposed that within access-based services consumers show less identification and personal attachment with the product. A consequence might be that product brands become less important. The information integration theory suggests that when new stimuli are becoming important, the importance of old stimuli will be reduced (Anderson 1981). Also according to this theory, service characteristics might become more important than brands in the context of access-based services. However, according to the theory of perceived ownership product brands still might remain important (Peck & Shu 2009). Consumers can become attached to a product by only touching objects, controlling or creating them, or by becoming familiar with an object (Belk 1988; Pierce, Kostova & Dirks 2001; Peck & Shu 2009). Product brands might still be important, since product identification may remain relevant for consumers. Baumeister, Scherer and Van Wangenheim (2015) studied the impact of product brands on ownership and access-based services. They found that product brands affect consumers their access attitudes less strongly in the car sharing context than while owning a product. Despite the reduced importance of product brands, it might still be important when making the decision whether a consumer wants to engage in car sharing. Besides, the magnitude of the importance of product brands is not investigated in the context of car sharing. This study will investigate how important the product brand is compared to the other attributes. In the past many studies were conducted in order to investigate the effect of price on consumer behavior (Armstrong & Overton 1971; Dodds, Monroe & Grewal 1991; Suk et al. 2010). The effect of pricing within access-based consumption is under-researched according to Lawson et al. (2016). Several studies found that consumers may have economic motives when engaging in car sharing (Lamberton & Rose 2012; Lawson et al. 2016). These studies inquire for further research to investigate the impact of price on the willingness to engage in access-based services. This is confirmed by Hildebrandt et al. (2015). These researchers indicate that the cost structure of car sharing needs to be investigated. This study will focus on the price per hour.
Garcia 2011; Choi & Oh 2010). Car sharing firms are also making use of electric and hybrid vehicles (Kley, Lerch & Dallinger 2011; Bohnsack, Pinkse & Kolk 2014). However, no research has been conducted to investigate whether consumers prefer electric, hybrid or conventional fuel vehicles in this context. Therefore, this study will focus on drive train as an attribute.
environmentally friendly than product ownership (Botsman & Rogers 2010; Martin, Shaheen & Lidicker 2010). However, Bardhi and Eckhardt (2012) did not find evidence that environmentally conscious consumers engage in access-based consumption. Therefore, the impact of environmental concern is investigated as a driver of engaging in car sharing.
Chen (2009) indicates that owning objects becomes less important. Nonetheless, Bardhi and Eckhardt (2012) suggest that consumers still prefer owning a car instead of sharing. Lawson et al. (2016) call for future inquiry to understand the relation between consumer materialism and access-based services. Since materialistic consumers prefer to purchase and own products, it is expected that consumer materialism is a barrier to make use of access-based services (Belk 1984; Belk 1985). Therefore, materialism may be a reason why consumers do not engage in car sharing. Bardhi and Eckhardt (2012) indicate that status will lead to a preference for owning objects. Owning a car contains more status than making use of car sharing (Bardhi & Eckhardt 2012). It might be the case that consumers their need for status reduces the likelihood that they engage in car sharing, since car sharing is associated with less exclusiveness (Lamberton & Rose 2012). Nonetheless, Lawson et al. (2016) found that status-oriented consumers are more likely to make use of car sharing. Since car sharing is less expensive than car ownership, these consumers are more likely to pretend that they are someone else for a day. This is easier for access-based consumption, since it is less expensive. This study will investigate the impact of consumers their need for status on engaging in car sharing.
2.3 Hypotheses
In this paragraph literature about the product attributes that affect the willingness to engage in car sharing will be discussed. Furthermore, the impact of the theoretically relevant moderators is being reviewed. In table 1 the literature review of the car sharing attributes is presented.
Table 1 -‐ Literature Review
Car brand Price per hour Drive train Reservation Vehicle station system
Baumeister, Scherer & Van
Wangenheim (2015) ü
Lamberton & Rose (2012) ü *
Lawson et al. (2016) ü *
Dagsvik et al. (2002) ü **
Hess, Train & Polak (2006) ü **
Hidrue et al. (2011) ü **
Horne, Jaccard & Tiedemann (2005)
ü **
Mabit & Fosgerau (2011) ü **
Hildebrandt et al. (2015) ü *** ü
* The studies only investigate the impact of price on engaging in access-‐based services, not specifically price per hour; ** The studies compared alternative drive trains to conventional vehicles in the context of car ownership;
2.3.1 Car Brand
In this study the effect of the following three car brand types on the willingness to engage in car sharing is investigated: prestige brands, volume brands and price brands (Desarbo & Manrai 1992; Kirmani, Sood & Bridges 1999; Park, Milberg & Lawson 1991). Prestige brands (e.g. Mercedes and BMW) contain excellent technical attributes and provide a high social status. They could be used to communicate a certain amount of status, wealth and exclusiveness (Richins 1994). Volume car brands (e.g. Ford and Toyota) are less expensive and are not primarily purchased because of their status or exclusiveness. They give value for money. Finally, price car brands (e.g. Dacia, Suzuki and Kia) are relatively inexpensive cars and do not provide status and exclusiveness. Price is an important factor for not purchasing and owning prestige brands. However, also the status that is derived from a brand and the technical performance are perceived as important (Saridakis & Baltas 2016). Lawson et al. (2016) indicate that consumers might feel the urge to pretend they are someone else for a day. This is easier for access than for ownership, since it is relatively inexpensive compared to ownership. Therefore, it might be the case that consumers will choose a prestige car brand instead of a volume or price brand. The exclusiveness of car sharing is perceived as low (Bardhi & Eckhardt 2012). However, it is still possible that the extent of exclusiveness within car sharing can vary by means of the product brand. Based on this, the study expects that prestige car brands are more preferred than volume and price car brands. This study also expects that volume car brands are more preferred than price car brands, since the quality of these cars is perceived to be better (Verhoef, Langerak & Donkers 2007).
Hypothesis 1a: A prestige car brand is expected to be more preferred than a price car brand. Hypothesis 1b: A volume car brand is expected to be more preferred than a price car brand.
2.3.2 Price per Hour
In the past lots of consumer studies found that price is one of the most important attributes when buying a product or using a service (Dawar & Parker 1994; Erickson & Johansson 1985; Moore 2004; Völckner 2008). Consumers that are willing to engage in car sharing may have financial motives (Bardhi & Eckhardt 2012; Lawson et al. 2016). Therefore, it is likely that price has a major impact on the willingness to engage in car sharing. Lamberton and Rose (2012) found that the price of a shared product has a negative influence on engaging in access-based services. They suggested that a lower-price per unit would draw people to participate in sharing programs. This study expects that a lower price per hour enhances the willingness to engage in car sharing.
2.3.3 Drive Train
Car sharing firms are making use of electric, hybrid and conventional fuel vehicles (Kley, Lerch & Dallinger 2011; Bohnsack, Pinkse & Kolk 2014). Electric vehicles offer lower operating costs than conventional fuel cars and are often recharged by power outlets (Eggers & Eggers 2011). The vehicles are efficient, do not emit pollution and can be driven almost noiseless. Hybrid vehicles emit in about fifty percent of the pollution a conventional car produces (Mabit & Fosgerau 2011). Several studies indicate that the consumer preferences for electric or hybrid cars depend primarily on the price of the vehicles (Lebeau et al. 2012; Potoglou & Kanaroglou 2007). If the price is high, it is less likely that consumers will make use of such a vehicle. Since the price for car sharing is relatively low, it is expected that consumers are more likely to make use of these vehicles in this case than while owning an electric or hybrid car. For electric vehicles also the driving range and recharging time of the battery are important (Eggers & Eggers 2011). However, electric vehicles have several downsides that could inhibit making use of these cars. The batteries that are used have a certain operating distance and recharging the batteries takes several hours. Besides, the performance depends on the outside temperature and the battery needs to be replaced after a certain amount of driven kilometers (Eggers & Eggers 2011).
Hypothesis 3a: An electric car is expected to be less preferred than a conventional fuel car. Hypothesis 3b: A hybrid car is expected to be more preferred than a conventional fuel car.
2.3.4 Reservation
Belk (2014) indicates that car sharing concepts differ with regard to the necessity to make a reservation or not. For some concepts it is not required to make a reservation, whereas for other car sharing concepts consumers need to make a short- or long-term reservation (Boyacı, Konstantinos & Geroliminis 2015). Hildebrandt et al. (2015) studied the impact of the technological features of reservation. They found that consumers were more likely to engage in car sharing if they were able to make an online reservation instead of using the phone. This attribute was perceived as important. However, little research has been conducted to the preferences for reservation type.
If the costs of engaging in access-based consumption are high, consumers will not make use of this consumption mode (Lamberton & Rose 2012). Based on this it is likely that when no reservation is needed, consumers are most likely to engage in car sharing. This offers the highest flexibility. However, it is possible that consumers want to be certain that a car could be used. Certainty also serves as an input for making a decision (Pham 1998). It may be inconvenient that another user is already making use of the car when a consumer wants to use it, since it was not an option to make a reservation. Nonetheless, the perceived costs of not making a reservation are lower than when making a reservation. Therefore, it is expected that making no reservation is preferred above making a long-term reservation. Boyacı, Konstantinos & Geroliminis (2015) further indicate that making a short-term reservation may be less effortful than a more long-term reservation. Reducing the perceived costs via a reduction in the effort to use the service may increase the attractiveness of car sharing (Lamberton & Rose 2012). Therefore, it is expected that a short-term reservation is preferred above the long-term reservation.
Hypothesis 4a: The possibility of consumers to not make a reservation is expected to be more preferred than the possibility to make a long-term reservation.
Hypothesis 4b: The possibility of consumers to make a short-term reservation is expected to be more preferred than the possibility to make a long-term reservation.
2.3.5 Vehicle Station System
system is the traditional variant of car sharing. The latter two variants emerged recently. Therefore, little research has been conducted to the consumer preferences for these vehicle station systems. Hildebrandt et al. (2015) conducted a conjoint analysis to car sharing preferences in which he used only two different vehicle station systems. The study only focused on the location where consumers have to pick-up their car. Firstly, they indicate that users might pick up a car at one or several stations. Secondly, they indicate that some car sharing cars are spatially dispersed in the city. Users need to locate these vehicles with their smartphone.
Hildebrandt et al. (2015) found that consumers want to access vehicles that are placed at one or more stations. A possible explanation is that consumers want to have certainty (Pham 1998). However, when consumers have the possibility to pick-up or drop-off the car at one station the user flexibility is low. Consumers have a limited opportunity to park nearby their destination. Since consumers want to have a certain amount of flexibility, it is expected that a fixed station will reduce the willingness to engage in car sharing (Bardhi, Eckhardt & Arnould 2012; Belk 2010). If consumers have the option to return the cars to multiple stations, the flexibility is increased. Therefore, this study expects that systems that allow users to return the cars to different stations are more preferred. Furthermore, the possibility to pick-up or drop-off the vehicle everywhere within the city offers consumers the flexibility to park wherever they want. Therefore, this concept is expected to be more preferred than returning the vehicle to the same location as where they picked it up. However, several studies indicate that parking is important for consumers (Button 2002; Grazi & Van den Bergh 2008). Due to the limited parking space it could be more desirable to park at a station of a car sharing firm. Despite these parking issues, this study expects that this system is more preferred than making use of a fixed station.
Hypothesis 5a: Returning the car to a different car sharing station as where consumers picked-up the car is expected to be more preferred than when they have to pick-up and return the vehicle to a fixed station.
Hypothesis 5b: Picking-up and returning the car anywhere within the city is expected to be more preferred than picking-up and returning the vehicle to a fixed station.
2.3.6 Moderators 2.3.6.1 Contagion
of disgust in which consumers are aware that other consumers physically touched an object. These researchers developed the theory of consumer-to-product contagion. Morales and Fitzsimons (2007) demonstrate that the product evaluations of consumers change in response to perceived physical contact of another consumer. When consumers think that another consumer has used a product, they could experience the feeling that the product is contaminated. This may lead to a negative product evaluation (Argo, Dahl & Morales 2008).
Bardhi and Eckhardt (2012) argue, in their exploratory research, that contagion might be a reason for consumers not to engage in car sharing. The impact of contagion on the use of access-based services might have important implications and declare the lack of utilization of car sharing in the Netherlands. Contagion typically manifests itself when the other person who uses the same service is unknown (Argo, Dahl & Morales 2006; Belk 2010). According to Bardhi and Eckhardt (2012) users of access-based services often do not know each other. Therefore, contagion is more likely to occur for access-based services. Bardhi and Eckhardt (2012) indicate that disgust might lead to a reduced sense of perceived ownership, as a result of which consumers are not willing to engage in access-based consumption modes. Touching, controlling, creating or becoming familiar with an object may lead to perceived ownership, as a result of which people develop a favorable attitude towards the object (Belk 1988; Pierce, Kostova & Dirks 2001; Peck & Shu 2009). A reduction in perceived ownership might lead to the development of unfavorable attitudes towards car sharing. Based on the theory of consumer-to-product contagion, this study expects that contagion will reduce the likelihood that consumers engage in car sharing.
Hypothesis 6a: Contagion decreases the probability that consumers engage in car sharing.
A lower price could be associated with a reduced product quality or a sloppy product (Rao & Monroe 1989). Therefore, consumers who are susceptible to contagion might prefer a higher price. However, contagion also may lead to a negative product or service evaluation. Therefore, car sharing will be less attractive to consumers. Consumers may become more price sensitive when contagion is present (Winkielman, Berridge & Wilbarger 2005; Morales & Fitzsimons 2007). Therefore, this study expects that contagion makes consumers more price sensitive.
Hypothesis 6b: Contagion amplifies the positive impact of the preference of a lower price per hour compared to a higher price per hour on the willingness to engage in car sharing.
2.3.6.2 Environmental Concern
friendly products. Consumers that have a pro-environmental orientation and make use of green products are seen as environmentally conscious (Lin & Chang 2012). These consumers are more likely to demonstrate green purchasing behavior. Access-based services, such as car sharing, are considered as more environmentally friendly than ownership (Botsman & Rogers 2010; Martin, Shaheen & Lidicker 2010). Bardhi and Eckhardt (2012) did not find evidence for the theory that consumers use access-based services to demonstrate pro-environmental behavior. However, Lawson et al. (2016) indicate that environmental concern may be an important motivation of consumers to use access-based services. Consumers that are interested in car sharing may believe that sharing offers a way to protect the environment or reduce waste, as a result of which they may derive a moral utility from sharing rather than owning (Minton & Rose 1997; Lamberton & Rose 2012). Therefore, this study hypothesizes that environmental concern increases the likelihood of engaging in car sharing.
Hypothesis 7a: Environmental concern increases the probability that consumers engage in car sharing.
Consumers that show environmental concern are more likely to demonstrate pro-environmental behavior (Hartmann & Apaolaza-Ibáñez 2012; Newton et al. 2015). Consumers that demonstrate pro-environmental behavior are more likely to choose an electric or a hybrid vehicle than a conventional fuel vehicle, since these consumers attach more value to the environmental aspects of products (Bockarjova & Steg 2014). This may also be the case when they are choosing for a drive train in the context of car sharing. Therefore, this study expects that environmental concern has a positive influence on electric vehicles and enhances the positive impact of hybrid vehicles on the willingness to engage in car sharing.
Hypothesis 7b: Environmental concern reduces the negative impact of the preference of electric vehicles compared to conventional fuel cars when engaging in car sharing.
Hypothesis 7c: Environmental concern amplifies the positive impact of the preference of hybrid vehicles above conventional fuel cars when engaging in car sharing.
2.3.6.3 Consumer Materialism
still might be more important than perceived ownership. Consumers may prefer possessing above sharing objects, as a result of which they are less likely to engage in car sharing. Therefore, this study expects that consumer materialism reduces the probability of engaging in car sharing. Hypothesis 8: Consumer materialism decreases the probability that consumers engage in car sharing.
2.3.6.4 Need for Status
Eastman, Goldsmith and Flynn (1999, p. 43) define status consumption as follows: “the motivational process by which individuals strive to improve their social standing through the conspicuous consumption of consumer products that confer or symbolize status for both the
individual and surrounding others.” Several studies indicate that consumers will remain
status-oriented (Prettenthaler & Steininger 1999; Bardhi & Eckhardt 2012). Status-status-oriented consumers use goods to display their social standing. Therefore, social influence has a huge impact on status consumption (Gao, Winterich & Zhang 2016). Schaefers, Lawson and Kukar-Kinney (2016) found that social influence is also present in the context of access-based services. They indicate that there is a positive effect of social influence on product ownership, since ownership is still associated with more status than sharing. Car sharing is associated with less exclusiveness than car ownership (Bardhi & Eckhardt 2012; Lamberton & Rose 2012). A consequence of this may be that status-oriented consumers are less likely to engage in car sharing. However, Lawson et al. (2016) found that consumers that have a high need for status are more likely to engage in access-based services. They indicate that consumers who are status-oriented might feel the urge to pretend they are someone else for a day. This is easier for access than for ownership. Based on this, need for status could also increase the likelihood to engage in car sharing. Nonetheless, it is expected that these status-oriented consumers are less likely to engage in car sharing.
Hypothesis 9a: Consumers’ need for status decreases the probability that consumers engage in car sharing.
This study expects that a prestige car brand is more preferred than a price car brand (hypothesis 1a). Consumers’ need for status might amplify this effect, since these brands are associated with more exclusiveness and status. Volume car brands are expected to be more preferred than price car brands (hypothesis 1b). Since volume car brands are more exclusive than price car brands, it is expected that consumers’ need for status will enhance this effect. Status-oriented consumers use consumption modes that are expensive. Therefore, it is likely that these consumers are less price sensitive (Frank et al. 2015; Baumeister, Scherer & Van Wangenheim 2015). This study expects that need for status reduces the price sensitivity of consumers.
Hypothesis 9b: Consumers’ need for status amplifies the positive impact of the preference of prestige car brands above price car brands when engaging in car sharing.
Hypothesis 9c: Consumers’ need for status amplifies the positive impact of the preference of volume car brands above price car brands when engaging in car sharing.
Hypothesis 9d: Consumers’ need for status reduces the positive impact of the preference of a lower price per hour compared to a higher price per hour on the willingness to engage in car sharing.
According to the cost-signaling theory, status-oriented consumers are interested in purchasing environmentally friendly products such as green cars (Griskevicius, Tybur & Van den Bergh 2010). This theory is based on the assumption that these consumers are willing to incur the costs for others. Noppers et al. (2014) also indicate that consumers purchase environmentally friendly products while having status motives. Griskevicius, Tybur and Van den Bergh (2010) suggest that when people consume their products in public, they are more likely to purchase environmentally friendly products. The car sharing service is consumed in public, as a result of which status-oriented consumers might make use of car sharing. Furthermore, they found that status-status-oriented consumers use green products if these products are more expensive. The price of the service could be seen as relatively low compared to owning a car. Therefore, the status of using an environmentally friendly vehicle on the willingness to engage in car sharing may be reduced. However, this study expects that need for status has a positive effect on the preference of electric vehicles and amplifies the impact of hybrid vehicles on the willingness to engage in car sharing. Hypothesis 9e: Consumers’ need for status reduces the negative impact of the preference of electric vehicles compared to conventional fuel cars when engaging in car sharing.
Hypothesis 9f: Consumers’ need for status amplifies the positive impact of the preference of hybrid vehicles above conventional fuel cars when engaging in car sharing.
2.3.7 Control Variables
Figure 1 -‐ Conceptual Model
Control Variables:
- Demographic variables (age,
gender, income and education)
- Number of kilometers driven
per week
- Car sharing knowledge
- Car ownership
- Car involvement
well-educated consumers are more likely to exert environmentally friendly behavior (Zander & Hamm 2012). Lawson et al. (2016) used age, gender, education and income as control variables. This study will also use a measure to see whether respondents have a certain amount of car sharing knowledge (Lamberton & Rose 2012). Beatty and Talpade (1994) indicate that product involvement is important for the usage of a product. Therefore, this study will control for car involvement. Meijkamp (1998) found that there is a negative relationship between engaging in car sharing services and car ownership. Therefore, the study controls for car ownership. Finally, the average number of kilometers driven per week will be considered. Consumers that use their car infrequently are more likely to make use of car sharing (Botsman & Rogers 2010). Figure 1 presents the conceptual model and summarizes the hypothesized relationships.
3. Methodology
This study applies a conjoint analysis. In this chapter the methodology of the study is discussed. Firstly, the conjoint measurement technique that is used for determining consumer preferences is described. Secondly, the estimation procedure will be discussed. Thirdly, the attribute levels are presented. Fourthly, the measures of the moderators are discussed. Fifthly, the choice design is reviewed. Finally, the utility function is being presented.
Price per hour* Car brand * Willingness to engage in car sharing Drive train* Reservation* Vehicle station system* Consumer materialism Contagion Environmental concern
Need for status
* The reference categories of the attributes are: price car brands, price per hour €12, conventional fuel vehicles, long-‐term reservation needed and pick-‐up and drop-‐off the vehicle at a fixed car sharing station.
8-3.1 Method of Preference Measurement
The conjoint analysis is the most popular method for measuring consumer preferences (Eggers & Sattler 2011). A conjoint analysis is a decompositional method. Participants evaluate the product attributes and levels jointly. The preference estimates of the product attributes are decomposed statistically. To analyze consumer preferences in the car sharing market, this study will rely on a choice-based conjoint analysis. In marketing research this technique has gained wide acceptance for identifying consumer preferences (Louviere & Woodworth 1983; Hennig-Thurau et al. 2007; Papies, Eggers & Wlömert 2011). A choice-based conjoint analysis leads to more accurate responses than a traditional rating- or ranking-based approach, since these analyses imitate a real-life purchase situation more accurately (Toubia, Hauser & Simester 2004). It is effective, since choices are an integral part of consumers’ daily life (Eggers & Sattler 2011). Another advantage of a choice-based conjoint is that a no-choice option could be included. This increases the realism of the study (Eggers & Sattler 2011).
3.2 Estimation Procedure
This study will estimate consumer preferences by making use of an aggregate Logit and a hierarchical Bayes conjoint analysis. The aggregate Logit model is used to study the consumer preferences on an aggregate level. The hierarchical Bayes conjoint analysis (HB) will be used in order to capture consumer heterogeneity (Andrews, Ansari & Currim 2004). By making use of this procedure, the reliability and predictive validity of the study is enhanced. According to Eggers and Sattler (2011) HB is the most widespread estimation procedure applied to choice-based conjoint data. It assumes that the preferences of participants are connected by a common multivariate normal distribution and offers a high level of detail.
3.3 Study Design
In this paragraph the attributes and the levels are discussed. Besides, the measures of the moderators are presented.
3.3.1 Attributes and Levels
levels are integrated in the conjoint analysis, the parameter estimates will become less reliable (Eggers & Sattler 2011). The attributes and their levels are presented in table 2.
The attributes and the levels follow from the hypotheses. The attribute car brand consists of three different levels: a prestige brand (Mercedes), a volume brand (Ford) and a price brand (Kia). In this study car sharing firms are not used as a brand, since the car sharing knowledge in the Netherlands is relatively low (Shaheen & Cohen 2013; Moeskops 2015). Therefore, it is expected that consumers do not have sufficient knowledge about these different car sharing brands. The attribute price per hour is subdivided into three price levels, ranging from a relatively low price per hour to a high price per hour. Since the cost structure of car sharing is complex, this study will only investigate the effect of price per hour. The attribute levels are based on actual prices in the Dutch car sharing market for hatchback cars (Ritjeweg 2016; Vereniging Voor Gedeeld Autogebruik 2016). Appendix A presents the cost structure of the most popular car sharing firms in the Netherlands. The price per hour varies from €2,50 till €15. In the Dutch market, car sharing firms make use of several drive trains. The attribute drive train is divided into three levels: a conventional fuel car, a hybrid car and an electric car. The attribute reservation consists of three levels: no reservation needed, a short-term reservation needed and a long-term reservation needed. The reservation types are based on the paper of Boyacı, Zografos and Geroliminis (2015). Furthermore, there are several vehicle station systems that vary in user flexibility (Boyacı, Konstantinos & Geroliminis 2015; Firnkorn & Shaheen 2016). Some systems require users to return the cars to the station where they picked-up the car, whereas others allow to return the cars to different stations. Finally, there are also systems that allow to take and leave the car anywhere within the city without strict vehicle stations. This attribute is divided into these different systems.
Table 2 -‐ Attributes and levels included in the conjoint study
Attribute Description Attribute Level 1 Attribute Level 2 Attribute Level 3 Car brand The brand of the car that is available at
the car sharing firm.
Kia Ford Mercedes
Price per hour
The price consumers have to pay to use the car for one hour.
€ 4 € 8 € 12
Drive train The drive train of the car. Conventional
gasoline/diesel car
Hybrid car Electric car
Reservation Deals with whether a short-term,
long-term or no reservation is needed to make use of car sharing.
No reservation needed 30 minutes before use (Short-term reservation needed) 12 hours before use (Long-term reservation needed) Vehicle station system
Some systems require users to return the vehicles to a fixed station, whereas others allow users to return the cars to multiple stations. Finally, there are also systems that allow users to pick-up and drop-off the car anywhere within the city without strict vehicle stations.
Return the car to a fixed station
Possibility to return the car to different stations
3.3.2 Measures
In this section the measures of the study are presented. All these measures will be rated on a seven-point Likert scale, ranging from strongly agree to strongly disagree. The measures of the control variables are shown in Appendix B.
3.3.2.1 Contagion
Existing literature does not provide a validated scale of consumer-to-product contagion. Therefore, this study will make use of existing theories of consumer-to-product contagion in order to develop a scale. Several studies indicate that consumers perceive a feeling of disgust when someone has touched the object (Mooy & Robben 2002; Peck & Childers 2003; Argo, Dahl & Morales 2006; Morales & Fitzsimons 2007; Argo, Dahl & Morales 2008). For example when consumers think that another consumer has tried a clothing item, they could experience the feeling that the product is contaminated. Also visually salient traces of previous users may lead to contagion, since consumers consider the product as contaminated (Argo, Dahl & Morales 2006). In figure 2, five items that will be used in this study are presented in order to measure consumers their susceptibility to contagion. By using a factor analysis this scale will be validated.
3.3.2.2 Environmental Concern
Environmental concern was measured by using a six-item scale developed by Kilbourne and Picket (2008). The measure could be seen as reliable in different contexts (Leonidou, Leonidou & Kvasova 2010; Bertrandias & Elgaaied-Gambier 2014; Polonsky, Kilbourne & Vocino 2014). The items are represented in figure 3.
3.3.2.3 Consumer Materialism
Consumer materialism was measured by using the six-item scale developed by Richins (1987). The measure could be seen as reliable in different contexts (Schroeder & Dugal 1995; Sirgy et al. 1998; Strizhakova & Coulter 2013). The items are represented in figure 4.
1. I do not mind trying on clothes that have been worn by others in the fitting room, if it is the only product left (reversed).
2. When opening a door, I try not to touch the door handle for too long.
3. I am sending back online orders when I see that the product packaging had been opened before.
4. I find using products that have been touched by others disgusting.
5. If I find fingerprints from previous users on a product, I won’t buy or use that product.
Figure 2 -‐ Items Consumer-‐To-‐Product Contagion Scale
1. I am very concerned about the environment. 2. Humans are severely abusing the environment.
3. I would be willing to reduce my consumption to help protect the environment. 4. Major political change is necessary to protect the natural environment. 5. Major social changes are necessary to protect the natural environment. 6. Anti-pollution laws should be enforced more strongly.
3.3.2.4 Need for Status
Need for status was measured by using the five-item scale of Eastman, Goldsmith and Flynn (1999). This measure could be seen as reliable. The scale is used to measure status consumption purchase intentions (O’cass & Frost 2002; O’cass & McEwen 2004; Fitzmaurice & Comegys 2006) and to determine the need for status (Han, Nunes & Drèze 2010; Lawson et al. 2016). The items are represented in figure 5.
3.4 Choice Design
This study will make use of a fractional factorial experimental design instead of a full factorial design. In the full factorial design all possible level combinations are shown to the respondents. In this study this would mean that the full factorial design contains 243 possible stimuli, since each attribute contains three levels (35). It is unreasonable to expose the respondents to such a high amount of stimuli. The complexity would even increase when stimuli are allocated across choice sets. In a fractional factorial design each respondent is only exposed to a subset of stimuli. According to Eggers & Sattler (2011) there are several criteria for selecting a choice design. Firstly, the attribute levels need to be balanced. All attribute levels should be displayed an equal number of times. Secondly, each attribute level pair should appear an equal number of times in order to be orthogonal. Thirdly, the alternatives in the choice set need to be maximally different from each other. Finally, there should not be dominating alternatives. The software package that is used automatically controlled for the first three efficiency criteria. However, it is difficult to control for the non-domination criterion since the preferences are unknown beforehand. In Appendix C an exemplary choice set is presented.
In each choice set three options are presented. In order to increase the realism of the study, a no choice option is included. There are two types of no-choice options: a dual-response no-choice option and a no-choice option as a separate alternative. A dual-response no-choice option will be used in this study to find out whether consumers are interested in car sharing, since this type
1. It is important to me to own really nice things. 2. I would like to be rich enough to buy anything I want.
3. It sometimes bothers me quite a bit that I cannot buy all the things I want. 4. I would be happier if I could afford to buy more things.
5. It is true that money can buy happiness.
6. People place too much emphasis on material things (reversed).
Figure 4 – Items Consumer Materialism Scale
Figure 5 -‐ Items Need for Status Scale
1. I would buy a product just because it contains status. 2. I am interested in new products with status.
contains more information than a no-choice option as a separate alternative (Wlömert & Eggers 2016). Furthermore, by making use of a real-life scenario the realism of the study is increased (the scenario is described in paragraph 4.1.2). The software package allows for random allocation of the stimuli to the choice sets. This study makes use of eleven choice sets and one holdout choice set (the holdout choice set is presented in Appendix D). The holdout choice set will be used as a benchmark. Eggers and Sattler (2011) indicate that a survey should include 12-15 choice sets. During the survey the respondents are provided with a short explanation about the attributes and levels. The explanatory text is available during the whole conjoint phase.
3.5 The Model
This study uses the random utility theory (RUT) as the conceptual basis for the model. The theory is based on the assumption that consumers choose the alternative that maximizes their utility (Jun & Park 1999). The RUT states that the utilities (U) of participants for products (or product attributes) are unobservable latent constructs (McFadden 1974; Manski 1977). The utility is divided into a systematic (V) and a random component (𝜀). The random component captures the (non-systematic) effects that are not accounted for. This leads to the following equation:
𝑈 = 𝑉 + 𝜀 (1)
The systematic component (V) links the product attributes (X) to their preference estimates (𝛽) in a choice-based conjoint study (Eggers & Eggers 2011). Equation 2 presents the aggregate Logit model, in which the utility (U) represents the utility of all respondents of the preferred car sharing option (j). This is the sum of all parameter estimates (𝛽) of the different attributes across the attribute levels (k). To measure the impact of the moderating effects, moderating variables will be created for this analysis.
𝑈!= 𝛽! 𝑋!" !
!!!
+ 𝜀! (2)
Furthermore, a multinomial Logit model (MNL) is used. MNL models the preferences in terms of choice probabilities (Swait & Louviere 1993). The likelihood that alternative a is chosen out of a set of J alternatives is:
𝑝 𝑎|𝐽 = exp (𝛽!∙ 𝑋!) exp (𝛽!∙ 𝑋!) !
!!!
(3)
𝑈!" = 𝛽!" 𝑋!" !
!!!
+ 𝜀!" (4)
The likelihood that respondent i will choose alternative a out of a set of J alternatives is:
𝑝! 𝑎|𝐽 = exp (𝛽!"∙ 𝑋!) exp (𝛽!"∙ 𝑋!) !
!!!
(5)
In the next section the equations that are presented are used. 4. Empirical Study
In this chapter the empirical study is reviewed. Firstly, the data collection and the procedure of the survey are discussed. Secondly, the sample of the respondents is described. Thirdly, a factor analysis will be conducted in order to reduce the dimensions of the moderators and control variables. Fourthly, the results of the aggregate Logit analysis are described. Finally, the outcomes of the hierarchical Bayes analysis are presented.
4.1 Choice Elicitation: Survey
4.1.1 Data Collection
For this study only respondents who live in the Netherlands (and are older than 18 years) were targeted to participate. In order to ensure that respondents understand the survey, a pre-test of the questionnaire was conducted. The think-aloud protocol was used to find out what comes up in the mind of a respondent. After the respondents (n=5) in the pre-test completed the survey, they answered questions about the content of the questionnaire (Simmons & Esser 2000). Firstly, it was asked whether the questions were formulated clearly. Secondly, questions were asked in order to determine whether the information provided about the attributes and levels was sufficient. Thirdly, respondents had to indicate whether the attributes and levels were considered as important. Finally, a question was asked in order to examine whether the importance of the attributes was reflected in the output. Based on the feedback of the respondents in the pre-test, the introduction and scenario were altered. The attribute vehicle station system led to some confusion. This definition was adjusted in order to clarify the meaning of this attribute.
4.1.2 Procedure
All respondents received a link of the online survey. The survey started with a brief explanation of the experiment on the introduction page. A short explanation of car sharing was provided to be sure that the concept of car sharing was clear to all respondents. Also an explanation of the attributes was used in order to let the participants become familiar with the attribute levels. The following scenario was being presented to the respondents:
[Please assume that you are in need for a car for a day every now and then in order to drive approximately 30 kilometers. Car sharing is available in all major cities in the Netherlands. If you currently own a car, please assume that this car does not work at the time you need it. You are considering a car sharing company to make use of a car. Below, you can find three alternatives to choose from. Please assume that the car sharing companies and cars do not differ apart from the attributes that are listed. For example, please assume that all car sharing companies are reliable, provide the same warranties, charge no membership fees, and no fees for kilometers driven (up to 100 km). All cars are less than one year old, have the same technical features, and are hatchbacks (e.g., Ford Focus, Mercedes A-class, or Kia Cee’d). Which of these options do you prefer?]
In the conjoint phase respondents were asked to choose the car sharing option that was most preferred for each choice set. During this phase, respondents were enabled to read the scenario again. Subsequently, the respondents indicated whether they would actually choose the preferred option. In this phase respondents could to click on a button in order to see what the attribute level meant. Respondents were randomly assigned to the choice sets. In the final section, questions with regard to the control variables and the moderating effects were asked. Also an attention test was used, in which consumers were asked to select number seven on a seven-point Likert scale. 4.2 Sample Characteristics
compared to the respondents that did not own a car (67,79%). Furthermore, the respondents were asked to indicate to what extent they are involved with cars on a seven-point Likert scale. Based on the means of the items, one can conclude that the respondents were moderately involved with cars. This is shown in figure 10. Finally, consumers were asked questions about their car sharing knowledge. Figure 11 indicates that the car sharing knowledge was relatively low.
Figure 9 – Number of Kilometers Driven per week
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 0-‐200 201-‐400 401-‐600 601-‐800 801-‐1000 N u m b er o f R esp on d en ts
Number of Kilometers Driven per week
0 10 20 30 40 50 60 70 80 90 100
No degree Secondary
School Community College (MBO) University of professional education (HBO) Bachelor’s
degree Master’s degree Philosophy Doctor of (PhD) N u m b er o f R esp on d en ts
Level of Education
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 <€30.000 Between €30.001 and €60.000 Between €60.001 and €90.000 Between €90.001 and €120.000 >€120.000 N u m b er o f R esp on d en ts Income Figure 7 -‐ Income 0 10 20 30 40 50 60 70 80 90 100 110 120 130 N u m b er o f R esp on d en ts Age Interval Age
Figure 6 -‐ Level of Education
Figure 8 -‐ Age
Figure 11 – Car Sharing Knowledge Figure 10 – Car Involvement
3,8 1,7 4,0 1,6 3,9 1,7 3,8 1,6 0 1 2 3 4 5 6 7
In general, I have strong interest in cars.
Cars are important to
me.
I get bored when people talk about cars.
My car means a lot to me.
Li ker t Scale (F ro m 1 to 7 ) Car Involvement Mean Std. Dev. 3,7 1,8 4,3 2,0 1,5 1,9 0 1 2 3 4 5 6 7
I am familiar with car
sharing programs. I have experience with car sharing programs.
I do not know much about how a car sharing program works. Li ker t Scale (F ro m 1 to 7 )
Car Sharing Knowledge
