The breakthrough of Mobile commerce

(1)

The breakthrough of Mobile commerce

Examining the determinants of adoption intention of mobile commerce

(2)

F.M. Hazelhoff – University of Groningen 2

The breakthrough of Mobile commerce

Examining the determinants of adoption intention of mobile commerce

Paterswolde, July 2012 Master thesis

Master of Business Administration

Department of marketing, Faculty of Economics and Business University of Groningen,

Frederik Hazelhoff

Student number: 1545515

Course program: Marketing Management & Market Research Email: frederik.hazelhoff@gmail.com

Address: Frederik Hazelhoff Patrijshof 3

9765 JD Paterswolde The Netherlands Supervision

(3)

Management summary

The aim of this study is to gain insights into the underlying factors explaining why consumers adopt to m-commerce. Most of the existing literature on m-commerce adoption tests simple models. In this study a comprehensive model is developed and tested to gain insights in the factors influencing m-commerce adoption. The conceptual model is developed by combining different adoption theories such as the technology acceptance model, theory of planned behavior and the theory of innovation adoption. First, we will test which determinants influence a person’s attitude towards m-commerce with a regression analysis. Secondly, a regression analysis is performed to test which determinants influence a consumer’s intention to adopt commerce. Finally, we test the mediation effect with attitude towards m-commerce as mediator.

The data which is used to test the proposed conceptual model is collected by performing an empirical study using a survey among Dutch consumers. The empirical research revealed that perceived usefulness and perceived ease of use are both important determinants of a person’s attitude towards m-commerce. These relationships are moderated by a person’s past experience with e-commerce and the innovativeness of a person. However, relative advantage, existing usage patterns, observability and testability have no significant effect on a person’s attitude towards m-commerce. A person’s attitude towards m-commerce, subjective norms and trusting beliefs in a mobile web vendor are the main factors influencing a consumer’s intention to adopt m-commerce. Perceived behavioral control appears not to influence a consumer’s adoption intention. Furthermore, perceived usefulness and perceived ease of use have an indirect effect on a consumer’s intention to adopt m-commerce via attitude.

(4)

Preface

This master thesis is my final work as student of the Master of Business administration Marketing Research and Marketing Management at the University of Groningen. With a particular interest in online marketing I was glad I could join the thesis group program with a common subject: Online marketing. The group environment was very stimulating and motivating in writing my thesis. I would like to thank some people in particular who helped me writing this thesis and throughout my study.

First of all, I would like that thank my parents for being supportive during my study. They always believed and encouraged me to finish my study and achieve a master degree in marketing.

Furthermore, I would also like to thank my supervisors Dr. Sonja Gensler and Alec Minnema for their support and constructive feedback which has been of great help in finalizing my thesis.

Finally, I would like to thank my friends and fellow students for their advice, help and great time during my study in Groningen.

(5)

1. Introduction

The last 10 years mobile phones have changed the way people live and work. Mobile phones do not only function as a communicative device nowadays, but also as a connective and transactional device. The possibilities of mobile phones have increased a lot, and more and more people own a smart phone with internet, cameras and location finders (GPS) on it. This has, for example, resulted in a large increase in the usage of mobile social networking. People want to stay connected and take part in social activities while they are on-the-go. When you look at social network sites like Facebook, LinkedIn and Twitter, you see a growth of at least 50 percent in their mobile audiences for the past year in the U.S.1 (table 1). Facebook has the largest mobile audience with more than 57 million mobile users.

Audience* for Selected Social Networking Brands

3 Month Avg. Ending Aug. 2011

Total U.S. Mobile Subscribers Ages 13+ (Smartphone and Non-Smartphone) *Includes mobile browser and app access.

Total audience (x1000)

aug-2010 aug-2011 % Change

Facebook 38240 57332 50%

Twitter 7639 13375 75%

LinkedIn 3234 5482 69%

Table 1: Facebook Mobile Audience Approaches 60 Million users

Another interesting change for companies and consumers is the ability to buy products and services online via your mobile phone. Mobile commerce (m-commerce) can be defined as any transaction with monetary value that is conducted via a mobile network (Clarke, 2001). It will allow consumers to purchase products and services over the internet without using a computer. Mobile commerce services include: Mobile advertising, mobile sales promotions, mobile entertainment, location-based mobile services, mobile browsing, mobile banking and mobile shopping (Barutçu, 2007). M-commerce was said to be the new e-commerce, building on the advantages of e-commerce (Dillard, 2010).

A few advantages of m-commerce over e-commerce are that it offers more ubiquity and accessibility to users (Schwiderski-Grosche and Knopse, 2002). Consumers can buy products and services anywhere, anytime.

1

(8)

F.M. Hazelhoff – University of Groningen 8 The ubiquity can be very interesting for time-critical applications like auctions, stock price changes and betting, where an alert notification can be sent. Another advantage is that vendors can know the location of their customers due to GPS technology and send promotions based on their location (Clarke, 2001). These way providers are able to offer new services or add value to current ones by taking the location into account (Barnes, 2003). A location-based service (LBS) is a software application which uses GPS to gain knowledge about where the mobile is located (Duri et al., 2001). A well-known example of LBS is Foursquare. If a person checks in at a certain location, they will earn points and this location will be shared with friends. Other examples are messages when there is a promotion in the neighborhood, finding location-based places like the closest ATM or people and location-based gaming. Furthermore, an important difference between m-commerce and e-commerce is the interface. Mobile screens are still a lot smaller than computer screens, which can lead to inconveniences during a purchase. However, nowadays screens are getting bigger and will be able to cope with higher resolutions taking away this constraint. Obviously, m-commerce has a lot of advantages over e-commerce, and mobiles have a great potential to play an important role in the buying process.

To make use of these advantages of m-commerce, companies have to understand how consumers search, evaluate and buy products and services using different platforms to stay ahead of their competitors. The way consumers shop has changed by the new possibilities of mobile phones. Last years, more and more people access the web via their mobile phone. Smartphones are, for example, often used to perform retail-related activities while they are in a brick-and-mortar store, bringing the digital and physical worlds together. While consumers are in the store, they can compare prices, locate the shop with the lowest price and read reviews of the products. Furthermore, extended information about the products can be received by scanning products and used to evaluate the products. For example, there is an application for iPhones and Androids, the GoodGuide mobile app, which allows you to check how safe, social responsible and how green a product is. At the moment, mobile phones are being mostly used in the information search and evaluation stages of the buying process but not in the final purchase stage.

(9)

F.M. Hazelhoff – University of Groningen 9 websites are becoming more suitable for web browsing and m-commerce. In 2011, 43% of the internet users in The Netherlands went online with their mobile phone compared to 21% in 2010 according to a research done by the CBS (2011). Based on this, an explosive growth in m-commerce was expected the past years. Larry Freed, president and CEO of Foresee Results said that mobiles are going to be a growing part of the interaction with companies, and it will only get bigger in the future.2 Even though, the percentage that is actually going to use m-commerce is low. According to a research from Forrester3 in 2010, only 4% from the Dutch participants ever bought a product with their mobile phone. Resistance towards innovations can be caused by conflicts in a consumer’s belief, or when an innovation requires substantial changes in behavior, leading to dissatisfaction (Ram and Sheth, 1989; Krackhardt, 1997). Barriers which are often mentioned for m-commerce resistance are inadequate security measures for mobile transactions, usability and small screens (Mahatanankoon and Vila-Ruiz, 2007). The barriers of m-commerce are well known. However, research on m-commerce adoption factors is rather scarce.

Research on the factors influencing the adoption of m-commerce exists. Maity (2010) suggests a model for m-commerce adoption based on qualitative research in the U.S. Hence, no conclusions can be made about the effect size of the antecedents. Wei et al. (2009) did research on the adoption of m-commerce in Malaysia. However, some important antecedents were not included such as behavioral attitude (Ajzen, 1991) and personal characteristics (Wei et al., 2009). This could lead to biased results due to omitted variables. A person’s attitude is defined as the overall evaluation of performing a certain behavior (Ajzen and Madden, 1986). Consumers with a positive attitude towards commerce are more likely to adopt m-commerce. Im et al., (2003) find that personal characteristics have an important influence on new-product adoption behavior. For example, when a person is more involved with mobile phones, and uses the latest models, it will be more likely he or she will also use his mobile phone for purchases. Wei et al. (2009) find a positive relationship between perceived usefulness, perceived cost (smartphone price, subscription fee, and service fee), social influences and a consumer’s intention to use m-commerce. The direct relationship between perceived ease of use and a consumer’s intention to use m-commerce is not found.

2

http://www.mobilecommercedaily.com ‘Apple store, Amazon offer best mobile retail experience: ForeSee Results’

(January 13, 2012)

3

(10)

F.M. Hazelhoff – University of Groningen 10 Yang (2005) studied the factors influencing m-commerce adoption in Singapore. In this study, the effect of individual characteristics on the attitude towards using m-commerce was explored. Perceived usefulness and perceived ease of use were added as moderators. Perceived usefulness and perceived ease of use are the two critical beliefs determining a users’ adoption intention and the actual usage according to the technology acceptance model (TAM) of Davis (1989). However, Pijpers et al. (2001) criticize that TAM studies fail to provide information on how consumers form a perception of new technologies and how these can be manipulated to increase the adoption of this technology. Combining the TAM with other models like the theory of planned behavior (TPB) (Ajzen, 1991) will overcome this problem.

The existing researches on m-commerce adoption mentioned above are either simple or incomplete. More research has to be done to gain insights into the underlying factors explaining why consumers adopt to m-commerce.

This research aims to identify, model and measure the relevant factors influencing m-commerce adoption by combining adoption models. Quantitative techniques will be used to test whether the sign and strength of the influence of the factors on the adoption of m-commerce correspond to the stated hypotheses. The research question is formulated as follows:

Which factors are influencing a consumer’s intention to adopt m-commerce?

Technology adoption models are widely used to examine the factors influencing a person’s attitude and behavioral intention toward using an innovation. Existing innovation adoption models will be discussed in the literature section. These existing models will be used to develop a model which is relevant for the adoption of m-commerce. Results from existing literature are mostly incomplete. This could lead to biased results due to omitted variables. Important factors influencing m-commerce adoption might not have been tested. In this study, a comprehensive model will be developed and tested. By combining the TAM (Davis, 1989), TPB (Ajzen, 1991) and the theory of innovation adoption (Rogers,1995) we will test for direct, indirect and moderating effects on a consumers intention to adopt m-commerce.

(11)

F.M. Hazelhoff – University of Groningen 11 using the results of this research, companies can focus on those antecedents that have a positive influence on the intention of use of m-commerce. The reasons why consumers adopt can be used in marketing campaigns to attract more m-commerce consumers.

(12)

2. Theoretical framework

In this chapter, we briefly review previous research on the determinants of adoption intentions towards new service or product innovations. Four theories on innovation diffusion will be discussed: Theory of Planned Behavior (TPB), Technology Acceptance Model (TAM) and the Theory of Innovation Adoption (TIA). The TPB and TAM are both extensions of the Theory of Reasoned Action (TRA) which will be discussed first, followed by a discussion about the link between these theories. All these models explore the positive factors influencing the decision to adopt. However, it might be interesting to investigate factors leading to innovation resistance as well, since these factors are typically not the same (e.g., Lennon et al., 2007; Lin, Hsin-Yu, and Sher, 2007). Finally, these theories will be combined and hypotheses will be presented regarding the factors influencing the adoption of m-commerce and a conceptual model will be presented showing the direction of the proposed effects.

2.1 Theories on innovation adoption

In literature, several theories can be found describing the factors influencing consumer behavior. Most research on technology acceptance concentrates on the following four models: Theory of Reasoned Action, Theory of Planned Behavior, Technology Acceptance Model, and the Theory of Innovation Adoption. Hundreds of studies can be found using one of these models to explain organizational or consumer acceptance of different ICT-systems and applications (Venkatesh and Davis, 2000). Each of these models will now be described in detail.

2.1.1. Theory of Reasoned Action (TRA)

The TRA was originally formulated in 1967 to examine the relationship between attitudes and behaviors (Fishbein and Ajzen, 1975). The model forms a simple basis which can be used for identifying where and how target consumers behavior change (Sheppard et al., 1988). According to the TRA, a person’s behavior is determined by his or her behavioral intention to perform a certain behavior (Fishbein and Ajzen, 1975). The behavioral intention can be measured as the strength of a person’s intention to perform a certain behavior (Fishbein and Ajzen, 1975). The intentions should capture the motivational factors influencing a certain behavior. How hard a person is willing to try or how much effort he or she wants to exert into performing the behavior in question.

(13)

F.M. Hazelhoff – University of Groningen 13 and subjective norm (see figure 1).

Figure 1: Theory of Reasoned Action (adopted from Fishbein and Ajzen 1975)

Attitude can be defined as an individual’s positive or negative evaluation of the performance effect of a particular behavior (Ajzen and Madden, 1986). Subjective norms refer to an individual’s perceptions of other people’s opinions on whether or not he or she should perform a particular behavior (Ajzen and Madden, 1986). The determinants attitude, subjective norm are both influenced by underlying believe structures.

Behavioral beliefs are the perceived likelihood of certain consequences of the behavior in question, weighted by an evaluation of the consequences, either positive or negative (Celuch and Dill, 2011). The normative believes are the perceived pressure from important referents, weighted by a person’s motivations to comply with these referents (Celuch and Dill, 2011). One of the assumptions of the TRA is that the behavior is voluntary and under control, however this appeared not to be one hundred percent sure in all cases (Ajzen, 1991).

2.1.2 Theory of Planned Behavior (TPB)

The Theory of Planned Behavior (Ajzen, 1991) is an intention model which is grounded in models from social psychology and is able to successfully predict and explains a wide range of behaviors (Taylor and Todd, 1995). The TPB extends the TRA (Ajzen and Fishbein, 1980; Fishbein and Ajzen, 1975) adding a third factor perceived behavioral control (PBC) for conditions where individuals do not have control over their behavior.

According to TPB an individual’s behavior can be explained by the person’s behavioral intention which is influenced by attitude, subjective norm and perceived behavioral control. The determinants attitude, subjective norm and perceived behavioral control are all influenced by underlying believe structures (see figure 2). Perceived behavioral control is defined as an individual’s perception of the presence or absence of the requisite resources or opportunities necessary for performing a behavior (Ajzen and Madden, 1986).

Behavioral Beliefs and Motivations Normative Beliefs and Motivation to Comply Behavioral

Intention Actual Behavior Attitude toward

Behavior

(14)

F.M. Hazelhoff – University of Groningen 14 Figure 2: The Theory of Planned Behavior (Adapted from Ajzen 1991)

Pavlou and Fygenson (2006) used the TPB to examine an internet user’s decision to purchase a product online. Results from this research showed that attitude and perceived behavioral control a significant part of the variance in adoption intentions, but subjective norms had no significant effect. Other research of Harrison et al. (1997) showed that all three TPB constructs predict unique variance in adoption intention. Attitude had the strongest effect on the adoption intention.

2.1.3 Technology Acceptance Model (TAM)

(15)

F.M. Hazelhoff – University of Groningen 15 Figure 3: Technology Acceptance Model (Adapted from Davis 1989)

The TAM replaces many of TRA’s attitude measures. According to Davis (1989) the perceived usefulness and the perceived ease of use are the most important determinants influencing the attitude toward using a technology. And usage intention is the sole determinant of actual system usage (see figure 3). The reason why this model is so successful is that system designers have some degree of control over the usefulness and ease of use of products or services. Perceived usefulness is defined as the degree to which a person believes that using a particular system would enhance his or her job performance. People might find a product or service useful, however, they believe that the product is hard to use and that the effort of using it out weight the performance benefits. Perceived ease of use is defined as the degree to which a person believes that using a particular system would be free of effort (Davis, 1989).

In the meta-analysis of Schepers and Wetzels (2007) the TAM relationships were confirmed. Both perceived usefulness and perceived ease-of-use showed significant effects on attitude and behavioral intention to use. It was also found that perceived usefulness has a stronger influence on the adoption of new technologies then the perceived ease-of use. According to Venkatesh and Davis (2000), numerous TAM studies explain typically around 40% of variance in usage intention and behavior. This indicates that perceived usefulness and perceived ease-of-use are important factors influencing the adoption of new technologies. However, it also suggests that extension of the TAM model is necessary to explain the adoption of new technologies. Researchers criticized that the generality of the TAM fails to provide useful information about the opinion of the specific system (Hu et al., 1999; Mathieson, 1991). Furthermore, Legris et al., (2003) concluded that the TAM model should be extended and include variables that relate to both human and social change processes. Therefore, in this study we include both subjective norms and perceived behavioral control as

External Variables Perceived Usefulness Perceived Ease of Use Attitute Towards Using Behavioral

(16)

F.M. Hazelhoff – University of Groningen 16 predictors of a consumer’s adoption intention to improve the predictability of the TAM model.

2.1.4 Theory of Innovation Adoption (TIA)

The theory of innovation adoption is another theory which provides valuable insights in the adoption of new technology (Rogers, 1995). The TIA has examined the adoption of new products and services from a diffusion of innovation perspective (Rogers, 1983; Tornatzky and Klein, 1982). Rogers proposed a model describing the process of decision making distinguishing five stages (See figure 4).

Figure 4: Innovation–Decision Process Model (adopted from Rogers 1995)

1. Knowledge: In this stage a person builds an understanding of the innovation and the functions it has. A person’s prior experience with corresponding innovation and personal characteristics will influence the knowledge formation. In this stage a person starts to build their beliefs about the innovation which will finally lead to a certain attitude towards an innovation.

2. Persuasion: In the second stage a person develops an opinion and attitude towards the innovation which could either be favorable or unfavorable. Characteristics of the innovation which will increase the adoption are (1) Relative advantage of the innovation, (2) Compatibility with the values and beliefs of the person, (3) The complexity, (4) Trialability and (5) Observability.

3. Decision: In the third decision stage a person will make a decision whether to adopt or reject the innovation based on the attitudes they formed in the previous stages.

(17)

F.M. Hazelhoff – University of Groningen 17 5. Confirmation: A person will finally reconsider the innovation adoption based on the

level of satisfaction on the usage experience of the innovation and will decide whether or not he or she will use the innovation in the future.

In this research the influence of individual characteristics and innovation characteristics on m-commerce adoption will be examined. The TIA is integrated in a more detailed framework of the determinants of m-commerce adoption.

2.1.5 The link between the theories on innovation adoption

The theories on innovation adoption discussed in the previous section are linked together. Both the TPB and the TAM are adaptations of the TRA model. Where the TIA has some overlap with the TAM model. An overview of the TRA, TAM, TPB and TIA can be found in figure 5.

Figure 5: Theory overview of the TRA, TAM, TPB and TIA.

Hartwick and Barki (1994) found in a study using the TRA that the impacts of subjective norm and attitude differ when the use of technology is voluntary or not. When the usage of a system is mandatory users would give greater weight to the opinions of others, since they think frequent use is appropriate. The intention of users under voluntary usage will be more based on their own attitudes (Hartwick and Barki, 1994). Therefore Ajzen (1991)

(18)

F.M. Hazelhoff – University of Groningen 18 extended the TRA model with the variable perceived behavioral control as third predictor of intention resulting in the TPB model.

The TAM model is another adaptation of the TRA which has been especially used trying to find the factors that influence an individual’s acceptance of IT (Davis, 1989). In the TAM the attitude measure are replaced by two technology acceptance measure usefulness and ease of use. Perceived ease of use and perceived usefulness are used as mediator of actual system use. In contrast with the TAM, attitude is measured by a person’s behavioral beliefs and the evaluations of the outcomes in the TRA. Furthermore, the main difference of the TPB and the TAM is that subjective norm is not proposed as antecedent of intention in the TAM. Davis et al. (1989) find that social influences had no significant effect on the intention to use and therefore not include in the TAM. However, it might be that subjective norms are an important driver for m-commerce adoption. Venkatesh and Davis (2000) find in an extension of the TAM, the TAM2, that subjective norms significantly influence user acceptance.

The TPB and TAM also differ in their treatment of behavioral control, which refers according to Mathieson (1991) to the opportunities, skills and resources needed to use a system. Perceived ease of use does incorporate internal control factors within the TAM but does not explicitly incorporates external control factors like time, opportunity and the cooperation of others. Because of it robustness, the TAM is less likely to identify idiosyncratic barriers. The TRA and TPB will be better in capturing situation specific factors since these models first identify control variables for every situation.

In Rogers (1995) diffusion of innovations perspective, five innovation characteristics serve as important antecedents to a person’s adoption decision: relative advantage, trialability, compatibility, complexity and observability. Complexity is as noted before defined closely to the definition of perceived ease of use (Rogers and Shoemaker, 1971). Perceived usefulness and relative advantage are highly correlated according to Plouffe et al. (2001). However, Youwei et al. (2011) find that relative advantage and perceived usefulness are related but distinct constructs. The main difference is that relative advantage fully mediates the effect of perceived usefulness of existing technology on the usage intention and partially mediates the effect of the perceived usefulness of the new technology.

2.2 Innovation resistance

(19)

F.M. Hazelhoff – University of Groningen 19 Bloem and Poiesz, 1997). However, stating that resistance is simply the opposite of adoption seems not appropriate (Gatignon and Robertson, 1989; Herbig and Day, 1992; Ram and Sheth, 1989). Garcia and Atkin (2002) claims that the factors defined in adoption research are typically not the factors which lead to active innovation resistance.

Therefore Kleijnen et al. (2009) developed a theoretical conceptualization of consumer’s resistance to innovations. Innovation resistance is defined as the resistance either caused by potential changes from a satisfactory status quo has to be made or because it conflicts with their belief structure (Ram and Sheth, 1989). Kleijnen et al. (2009) makes a distinction between seven drivers of consumer resistance. These can be split into two main groups. The first group of drivers exists when innovations require a change in consumer’s established behavioral patterns, norms, habits and traditions which are likely to be resisted. Second, innovations may cause a psychological conflict or problem for consumers and are likely to be resisted.

Compatibility is the extent to which new products or services are consistent and compatible with consumer’s needs, beliefs, value, experiences and habits (Rogers, 1962). This seems close to the definition of the first group of resistance drivers. However, the operationalization of compatibility is inconsistent across studies. Therefore, Kleijnen et al. (2009) made a distinction between conflicts with traditions and norms and conflicts with existing usage patterns. Whereas tradition and norms can create a barrier if a behavior is contrary to group norms, or societal and family values (Herbig and Day, 1992). Sheth (1981) find that resistance can be a consequence of habits. When an innovation conflicts with the usage patterns of competing products, or contradict with workflows, practices or habits are more likely to face resistance (Hurter and Rubenstein, 1987).

(20)

F.M. Hazelhoff – University of Groningen 20 1975). Consumers often face uncertainties about the adoption of innovations, especially about the performance of the innovation and whether the outcomes of using the innovation are positive or negative (Garcia and Atkin, 2002). Several forms of risks have been mentioned in relation to consumer resistance (Bredahl, 2001; Ram and Sheth, 1989). Physical risk refers to the possible damage to the person or property may be caused by an innovation (Klerck and Sweeney, 2007). Economic risk is related to the costs of an innovation (Kleijnen et al., 2009). Functional risk is related with the performance uncertainty and social risk with concerns that the purchase or use of an innovation will not be approved by relevant others (Ram and Sheth, 1989).

By looking at the drivers of consumer’s resistance, companies can develop strategies that help overcoming these barriers which will finally lead to innovation adoption. For example, Kleijnen et al. (2009) find that a price-skimming strategy will lead to a postponement of innovations, since economic risk is one of the main drivers of adoption postponement.

Therefore, some of the antecedents of consumer resistance will be used in developing a framework exploring the factors which will lead to m-commerce adoption. As mentioned, the operationalization of the concept of compatibility is inconsistent across studies. Therefore we choose to use the definition of Kleijnen et al. (2009). The definition of tradition and norms are closely related to the definition subjective norms as defined by Ajzen and Madden (1986). Since the relationship between subjective norms and a consumers intention to adopt is well established (e.g. Shin, 2007; Nasco et al., 2008; Wei et al., 2009) we choose to include subjective norms instead of traditions and norms in the model. Existing usage patterns will be included used instead of compatibility. Furthermore, social risk will be added to the model as moderator of the effect of observability on a consumer’s attitude towards m-commerce. When m-commerce is not accepted by the group it is expected that consumers are less likely to show to others they are using their phone for m-commerce, negatively moderating the effect of observability on attitude. The other drivers of consumer resistance, perceived image, information overload, physical risk, economic risk are more applicable on products than on services. Since m-commerce is a service we choose not to include these drivers in our model. 2.3 Theories on innovation adoption and resistance applied on m-commerce

(21)

F.M. Hazelhoff – University of Groningen 21 theories a new model is proposed in this research to explore the factors influencing the adoption of m-commerce.

Perceived usefulness

When consumers expect that using a service will improve his or her job performance or increase the effectiveness and efficiency of daily activities it is more likely that consumers will intend to use this service (Davis, 1989). In system literature, the usefulness of a technology or device has a positive impact on the users’ intention to use the technology in the future (Rao & Troshani, 2007). The usefulness of m-commerce is strongly driven by the advantages of mobile services which include ubiquity, localization and timeliness (Wong and Hiew, 2005). Therefore, in this research the perceived usefulness will be defined as the belief that the use of m-commerce will improve a person’s job performance and daily activities. Hypothesis 1a: Perceived usefulness positively influences the attitude towards the usage of

m-commerce.

Perceived ease of use

Increasing the ease of use, convenience of buying products or services with a mobile phone will increase the adoption of m-commerce. With the small screens of mobile phones consumers might have trouble browsing through websites finding the right products. Therefore, companies should make their website applicable to all smartphones. According to Rogers (1995), complexity of systems will discourage consumer’s intention to use an innovation. Furthermore, extensive research exists providing evidence of the positive influence of the perceived ease-of-use on the attitude towards the usage of a technology (Hu et al., 1999; Jackson et al., 1997; Rogers, 2003; Venkatesh & Davis, 2000). Tornatzky and Klein (1982) find that when the degree of complexity is low a person is more likely to form a positive attitude about the innovation. Complexity is defined as the degree to which people see the technology as difficult to understand and use which is almost the same as the definition of perceived ease of use.

Hypothesis 1b: Perceived ease of use positively influences the attitude towards the usage of m-commerce.

(22)

F.M. Hazelhoff – University of Groningen 22 According to Rogers (1962), relative advantage is the degree to which consumers see a product or service as different and better then substitutes. In the case of m-commerce the relative advantages over e-commerce are ubiquity and accessibility (Schwiderski-Grosche and Knopse, 2002). A relative disadvantage is the reliability and security concerns for mobile commerce technology (Siau and Shen, 2003). As mentioned before relative advantage and perceived usefulness are related but measured differently. Since m-commerce was said to be the new e-commerce it will be interesting to investigate whether these advantages create a positive attitude towards m-commerce.

Hypothesis 1c: Perceived relative advantage of m-commerce over e-commerce positively influences the attitude towards the usage of m-commerce.

Compatibility

According to Rogers (2003) the compatibility of the innovation is one of the characteristics that will form a person’s attitude towards an innovation. Compatibility is the extent to which new products or services are consistent and compatible with consumer’s needs, beliefs, value, experiences and habits (Rogers, 1962). Kleijnen et al. (2009) made a distinction between the traditions and norms, which relate to a societally-relevant context and existing usage patterns, which refer to the personal routines and habits of individual consumers. Conflicts with tradition and norms can for example lead to highly undesirable consequences for the society (Saba et al., 2000). This is actually not the case with m-commerce since it only effects your own behavior. The focus in this research will therefore be whether or not the use of m-commerce conflicts with existing usage patterns. When consumers are satisfied with their current situation, in this case e-commerce, they will have less desire or reason to change to m-commerce (Foxall, 1993; Foxall, 1994; Ram, 1987; Sheth, 1981). Innovation that conflicts with existing usage patterns like well-established workflows, practices or habits will more likely face resistance. Therefore we hypothesize: Hypothesis 1d: Existing usage patterns negatively influences the attitude towards the usage

of m-commerce. Observability

(23)

F.M. Hazelhoff – University of Groningen 23 amount of positive or negative reactions (Rogers, 1995). Consumers carry their mobile phones everywhere, also when they are shopping in stores. This might create awareness of mobile phones usage for shopping influencing the formation of an attitude towards m-commerce usage.

Hypothesis 1e: Observability positively influences the attitude towards the usage of m-commerce.

Trialability

Rogers (2003) points out that consumer’s resistance toward an innovation can be overcome when consumers have to opportunity to try an innovation for a certain period of time. Trialability is described as whether the product can be tested or experimented with on a limited basis, this will result is less uncertainty (Rogers, 2003). When a consumer purchases a product using m-commerce he or she can evaluate whether or not it was success. If the purchase was no success they can always return to e-commerce or purchase the product in-store. Several studies on technology adoption found a positive relation between trialability and the adoption of specific technologies (Karahanna et al., 1999; Chau and Hu, 2001).

Hypothesis 1f: Perceived trialability positively influences the attitude towards the usage of m-commerce.

Personal Characteristics

Research on the implementation success of information technology showed that the success depends as much on individual characteristics as on the technology itself (Harrison and Rainer, 1992; Nelson, 1990). According to Rogers (1995), personal characteristics of the decision maker will have an influence on whether he or she will be persuaded to the next stage of the decision making process. Information of the knowledge stage can be helpful to decide which groups should be targeted and which should not.

(24)

F.M. Hazelhoff – University of Groningen 24 Figure 6: Innovation theory of Rogers (1995): product lifecycle of an innovation

To profile innovators, personal characteristics like age, social participation, income, education, innovativeness and past experience can be used according to Im et al. (2003). Innovators are those that have a higher levels education, income, leadership, have favorable attitude towards risk and are younger (Rogers, 1995).

Age has been linked a lot in adoption studies. Kerschner and Chelsvig (1981) find for example, that there was a negative relation between age and innovation adoption of ATMs, cable television, video games and video recorders. Older people are said to be more resistant to technological change then younger people, while younger people tend to pursue innovativeness (Rogers, 1995). Kolodinsky et al. (2004) find that older people are less likely to adopt to phone and PC banking then younger people. Testing possibilities and demonstrations might help these older people to gain comfort and confidence in new technologies. M-commerce is a new technology, and therefore expected that older people will find more difficulties to adopt to this technology and stick with the older technology like e-commerce. The change people have to make, the risks and new required skills to use m-commerce will resist older people more than younger people.

Hypothesis 2a: Age negatively influences the relationship between perceived usefulness, perceived ease of use, relative advantages, existing usage patterns and the attitude towards the usage of m-commerce.

(25)

F.M. Hazelhoff – University of Groningen 25 high. For consumers with higher income these costs will be less of a problem and adopt more quickly to m-commerce.

Hypothesis 2b: Income positively influences the relationship between perceived usefulness, perceived ease of use, relative advantages, existing usage patterns and the attitude towards the usage of m-commerce.

Hypothesis 2c: Education positively influences the relationship between perceived

usefulness, perceived ease of use, relative advantages, existing usage patterns and the attitude towards the usage of m-commerce.

In prior research it is believed that past experience with using a similar technology contributes greatly favorable towards new technology adoption (Dabholkar, 1996; Dickerson and Gentry, 1983). In the context of this research, people with experience with e-commerce are more likely to adopt m-commerce. The past experience with e-commerce has an impact on a person’s beliefs and attitude towards m-commerce as well (Levin and Gordon, 1989). People with past experience will most likely be more skillful and used to these technologies. Hypothesis 2d: Past experience with e-commerce positively influences the relationship

between perceived usefulness, perceived ease of use, relative advantages, existing usage patterns and the attitude towards the usage of m-commerce. Innovative individuals are those that are seeking information about new ideas. They are prone to risks and exhibit a low uncertainty avoidance index, and develop positive intentions towards acceptances (Rogers, 1995). M-commerce is a relatively new service with great advantages in comparison to e-commerce and brings some risks with it.

Hypothesis 2e: Innovativeness positively influences the relationship between perceived usefulness, perceived ease of use, relative advantages, existing usage patterns and the attitude towards the usage of m-commerce.

Social risk

(26)

F.M. Hazelhoff – University of Groningen 26 by their social environment (Kleijnen et al., 2009). Social risk is closely related to observability as defined by Rogers (2003). The observation of peer groups are important in the decision making process of consumers. A lack of social support of an unaccepted innovation could isolate users from their social system (Kleijnen et al., 2009). The effect of observability on attitude can be moderated by social risk. If a person does not want others to know that they use their smartphone during shopping it might lead to a more negative attitude towards m-commerce. A lack of social support of an innovation can eventually lead to isolation from their social system (Kleijnen et al., 2009).

Hypothesis 2f: Social risks negatively influences the relationship between observability and attitude towards the usage of m-commerce.

Attitude

Attitude is one of the main antecedents of intention to use in the TRA, TPB and TAM. High values of attitude indicate a positive attitude towards an innovation and are therefore more likely to adopt the technology (Nasco et al., 2008). The positive relationship between attitude and predicting adoption intentions has been well established in marketing literature (Harrison et al., 1997; Pavlou and Fygenson, 2006; Venkatesh et al., 2003).

A behavioral belief can for example be that m-commerce saves time in comparison to e-commerce. The evaluation of this outcome is determined whether or not it is desirable. If a consumer sees this outcome as favorable and desirable the intention of using it will increase. For instance, if a person belief that m-commerce will increase business transactions and he or she evaluates this as desirable, the intention to use m-commerce will increase.

Hypothesis 3a: Positive attitudes towards m-commerce positively influence a consumer’s intention to adopt m-commerce.

Subjective norm

(27)

F.M. Hazelhoff – University of Groningen 27 Interpersonal influences include friends, family, superiors and social networks such as peers. Mass media includes television, radio, internet and newspapers. Previous research on the effect of subjective norm on the adoption of m-commerce showed a positive relation (Shin, 2007; Nasco et al., 2008; Wei et al., 2009). Research of Carrol et al. (2002) find that persons that do not use mobile technology, like chatting, SMS or e-mailing, appear to struggle to maintain their social links. Therefore, using mobile services like m-commerce can be a way to maintain membership or increase social interaction (Se-Joon and Kar Yan, 2006).

Hypothesis 3b: Subjective norms positively influence a consumer’s intention to adopt m-commerce.

Perceived behavioral control (PBC)

In a study of the comparison between the TRA and TPB it is found that including perceived behavioral control enhances the prediction of behavioral intention and behavior (Madden et al., 1992). Perceived behavioral control will therefore be added as variable influencing the adoption of m-commerce. If a person feels he or she has control over a performing a certain behavior the person is more likely to form strong intentions to perform the behavior (Notani, 1998). The more resources and opportunities a person beliefs he or she has the more perceived behavioral control they have over the behavior (Madden et al., 1992). A person can for example believe that he or she does not have the knowledge to use m-commerce. And that knowledge is an important determinant whether or not a person intends to adopt m-commerce. People’s behavior is heavily influenced by their confidence in the ability to perform that behavior. Their self-efficacy beliefs can have great influence on a person’s choice of activities, preparation and emotional reactions on a behavior (see Bandura, 1982, 1991).

In the study of Chau and Hu (2001) there was found a significant positive influence of PBC on behavioral intentions. Other studies showed significant positive effects of PBC on the intention to use a certain technology (Chang, 1998; Mathieson, 1991; Venkatesh et al., 2003). Results from Chang’s (1998) study showed that PBC was the most important predictor of intention to use illegal software.

(28)

Trusting beliefs in a mobile web vendor

A big issue concerning m-commerce adoption is trust. Trust can be regarded as the willingness of a party to be vulnerable to the actions of another party based on the expectations that the other party will perform particular actions (Mayer et al., 1995). The difference with trust in an offline setting is that online trust is generated with the interaction of an online information system. When the first transaction with a mobile phone is a disappointment, there is a high chance of drop out due to distrust. Consumers will be suspicious of its ability to deliver on its promises (Siau and Shen, 2003; Schwiderski-Grosche and Knopse, 2002). With m-commerce consumers do not meet face-to-face and are often concerned whether their money and personal information is safely transferred to the third party (Luarn and Lin, 2005). Consumers also have concerns about whether the company makes good on its side of the deal and behave as promised. According to Jarvenpaa and Tractinsky (1999) trust has a direct effect on a consumer’s purchase intentions. Wu et al. (2011) find in a meta-analysis a strongly positive relation between trust and the adoption of new technology. For online transactions like m-commerce trust reduce the perceived social complexity of communicating with m-commerce vendors (Gefen and Straub, 2003).

Hypothesis 3d: Trusting beliefs in a mobile web vendor positively influence a consumer’s intention to adopt m-commerce.

2.4 Conceptual model

(29)

F.M. Hazelhoff – University of Groningen 29 innovation characteristics that are important in forming an attitude towards an innovation should not be ignored when investigating adoption intentions of innovation. Therefore Relative advantage, existing usage patterns, observability and trialability are added as predictors of a person’s attitude towards m-commerce. Furthermore, opinions of others are important in the decision making process of m-commerce adoption. If m-commerce is generally accepted by the group a person belongs to, e.g. if the group belief m-commerce is safe and useful, the barrier to use m-commerce is expected to lower. Finally, it is expected that when the mobile web vendor can be trusted and keep their information safe a consumer is more likely to use m-commerce. In this study not only the direct factors on a consumer’s adoption intentions are investigated but also the indirect factors via a consumer’s attitude towards m-commerce. A visual representation of the theoretical constructs can be found in figure 7.

TAM

TIA

TPB

Figure 7: Conceptual model

H1a) Perceived usefulness (+)

H1c) Relative advantage (+) H1b) Perceived ease of use (+)

H1d) Existing usage patterns (-)

H1f) Trialability (+) H1e) Observability (+) H2f) Social risk (-) Moderators: H2a) Age (-) H2b) Income (+) H2c) Education (+) H2d) Past experience (+) H2e) Innovativeness (+) Consumer intentions to adopt m-commerce H3a) Behavioral attitude (+)

H3b) Subjective norms (+) H3c) Perceived behavioral control (+)

(30)

3. Methodology

This chapter elaborates on the research method that is used to test the hypotheses and finally answer the research question. It includes a specification of the data collection and the sample, the measures used and finally the plan of analysis.

3.1 Research Method

The method that is used to answer the research question is of descriptive nature and a quantitative research design is used to test the hypotheses. The purpose of this research is to find possible antecedents of m-commerce adoption. To empirically test the proposed hypotheses, a multiple regression analysis will be applied. Regression analysis is used to statistically test relationships between a metric dependent variable and multiple independent variables (Malhotra, 2007).

3.2 Data collection and the sample

To obtain the data for this research a survey method is used. Respondents are asked to fill in an online questionnaire. The survey is in Dutch and is distributed via Facebook, family, friends and other master students via e-mail. The respondents are also asked to forward the survey to others people to collect more data. An online survey is used because it is a fast way to collect data and it will be easy to process the data once collected. Participation is voluntary and anonymous with no rewards and the survey is the same for every respondent.

The target group for this research consists of males and females between 17 and 65 years. The only pre-condition is that the participants should have a smartphone. The participants should have experience using their mobile phone for browsing the internet to be able to form an opinion regarding the questions about m-commerce. Unfortunately, not all the respondents completed the online questionnaire. Out of the 245 participants, only 194 completed the questionnaire (79%). Possible reasons for the nonresponse error are inability and unwillingness errors. For the purpose of this research, which is finding associative relationships, the sample size is large enough. In similar studies on m-commerce adoption sample sizes of approximately 200 respondents were used (e.g. Wei et al., 2009; Yang, 2005).

(31)

F.M. Hazelhoff – University of Groningen 31 3.3 Variable measurement

As can be seen in the conceptual model (figure 7) the dependent variable in this research is the consumer’s adoption intention of m-commerce. The independent variables, which are tested whether they explain significant variation in the dependent variable, can be divided into two groups. The first group consists of the four variables that have a direct effect on the adoption intention and the second group consists of six variables with an indirect effect with attitude as mediator. The question about trialability of m-commerce is only asked to respondents who had ever bought a product with m-commerce. To be able to judge about the trialability of m-commerce a respondent should have actually used m-commerce.

All the constructs, except the demographic variables are measured with 1-7 Likert scales. With a score of 1 indicating that a respondent disagrees with a statement and a score of 7 that he or she agrees with the statement. Likert scales are understandable for respondents and can easily be used in performing statistically analysis. The Likert scales can be considered as ordinal scales, however in this research we assume that these scales can be considered as interval scales and therefore be used for regression and factor analysis (Malhotra, 2007). Most of the items that are used to measure the constructs of the dependent and independent variables are based on previous scientific studies. Some of these items are adapted to suit in the context of m-commerce. An overview of the items used and its sources can be found in Appendix B.

(32)

F.M. Hazelhoff – University of Groningen 32 occur if the independent variables correlate with the moderators (Aiken and West, 1991). 3.4 Plan of analysis

The analyses that are performed are discussed in the next chapter. First, a reliability analysis is performed to test the reliability of the different constructs.

In total 43 questionnaire items are used to measure 19 constructs. To check for internal consistency of the constructs a Cronbach’s Alpha reliability analysis is performed. After the reliability analysis, a factor analysis is performed to see if the items that highly correlate may be used as independent variables in the regression analysis. When these analyses are completed it can be concluded if the items can be averaged into the different constructs.

Secondly, three regression analyses are performed to test the direct effects on attitude and adoption intention. The effect of trialability on attitude is tested separately since in this sample only 57 respondents actually tried m-commerce before. Small sample sizes might result in too little statistical power for the test to realistically identify significant results (Hair, 2010). Therefore the effect of perceived usefulness, perceived ease of use, relative advantage, existing usage patterns and observability on attitude is tested in a separate regression analysis. Before the actual model estimation, it is tested whether the assumptions of OLS regression are violated. The models are tested on linearity, non-normality, heteroscedasticity, and multicollinearity.

(33)

4. Results

In this chapter the results of the analysis are presented. First, the demographic profile of the respondents are discussed, followed by a discussion about the reliability and validity of the sample. Then, the parameter estimates for the models will be estimated. Finally a moderator and mediator analysis is performed.

4.1 Participants

The total sample of the online survey consists of 194 respondents. An overview of the demographic profile of the respondents can be found in table 2. The sample consists of 59.3% males and 40.7% females. 20.0% of the respondents did not want to answer their income. Of the remaining participants, 51.9% have a monthly income of less than €500 or between €500 and €1500. Most of the respondents, 42.3%, have a Mavo or VMBO certificate. The dominant age group in this sample is 20 years or younger (44.3%).

Variables Frequency Percentage Variables Frequency Percentage

Gender Male 115 59.3 Age ≤ 20 years 86 44.3

Female 79 40.7 21-25 years 46 23.7 Income < €500 42 26.9 26-30 years 20 10.3 €500 - €1500 39 25.0 31-35 years 7 3.6 €1500 - €2500 30 19.2 36 - 40 years 14 7.2 €2500 - €3500 24 15.4 ≥ 41 years 21 10.9 > 3500 21 13.5 No answer 38

Education Primary School 2 1.0

Mavo/VMBO 82 42.3

Havo/VWO 29 14.9

HBO 49 25.3

WO 32 16.5

Table 2: Demographic profile for respondents

(34)

F.M. Hazelhoff – University of Groningen 34 28,4% 20,6% 3,1% 28,4% 10,8% 51,0% 0% 10% 20% 30% 40% 50% 60% Compare prices Review products Review shop Finding closest shop Buying products Not using

Smartphone usage while shopping

28,4% 20,6% 3,1% 28,4% 10,8% 51,0% 0% 10% 20% 30% 40% 50% 60% Compare prices Review products Review shop Finding closest shop Buying products Not using

Smartphone usage while shopping

28,4% 20,6% 3,1% 28,4% 10,8% 51,0% 0% 10% 20% 30% 40% 50% 60% Compare prices Review products Review shop Finding closest shop Buying products Not using

Smartphone usage while shopping

smartphone for a longer time are more likely to buy a product or service using m-commerce. Smartphone

possession Frequency Percentage

< 6 months 15 26.3

6 - 12 months 4 7.0

1 - 2 years 13 22.8

> 2 years 25 43.9

Table 3: Smartphone possession duration

Value Df Sig. (2-sided) Pearson Chi-Square 9.910 3 0.019 Table 4: Chi-Square test

The respondents were also asked a question how they used their smartphone during shopping. Most of the respondents, 51%, did not use their smartphone while shopping (see figure 8). Comparing prices (28.4%) and finding the closest shop (28.4%) were the most used features of a smartphone.

Figure 8: Smartphone usage while shopping

4.2 Scale reliability and factor analysis

In this research multiple questions are asked in the online questionnaire to measure one construct. If we want to combine these items into one construct we need to test its reliability. The Cronbach’s alpha is the most common used statistic to test the internal reliability of a set of items. According to Malhotra (2007) a Cronbach’s alpha of below 0.6

(35)

F.M. Hazelhoff – University of Groningen 35 indicates unsatisfactory internal consistency reliability. In table 5 the results of the reliability analysis can be found. Three constructs have a Cronbach’s alpha of below 0.6. For existing usage patterns one item is excluded increasing the alpha from 0.147 to 0.517, which is still below 0.6. Excluding another item did not increase the alpha any further. We choose to delete item 1 for the existing usage patterns construct. For the other constructs with an alpha of less than 0.6 we choose to keep the original scales. Leaving out 1 item would result in single item scales. A disadvantage of single item scales is that they show only a small degree of differentiation and estimation of reliability is not possible. Furthermore, to be able to compare the results with other researches we choose to keep the original scales.

Constructs Items Initial Cronbach's Alpha

Final Cronbach's Alpha

Subjective Norms 3 0.87 0.87

Trusting beliefs in web vendor 3 0.859 0.859

Perceived Ease of Use 3 0.853 0.853

Perceived Usefulness 4 0.846 0.846

Perceived Behavioral Control 3 0.755 0.755

Attitude 3 0.704 0.704

Observability 2 0.586 0.586

Relative Advantage 2 0.498 0.498

Existing Usage Patterns 3 0.147 0.517 (item 1 excluded) Table 5: Cronbach’s alpha, internal consistency

After the reliability analysis a factor analysis is performed for both models to test whether the items that highly correlate may be combined into one factor or independent variable. This is done because the two models have different dependent variables. The first factor analysis will include the items that expected to have an effect on attitude and in the second factor analysis the items are included that are expected to have an effect on a consumer’s intention to adopt m-commerce

To test if a factor analysis is appropriate two tests are performed; the Kaiser-Meyer-Olkin (KMO) measure and the Bartlett’s test of sphericity. The Bartlett’s test is significant (p=0.000), rejecting the null hypothesis, stating that the population correlation matrix is an identity matrix (Appendix C, table 1). According to Malhotra (2007) KMO values of below 0.5 indicate that factor analysis may not be appropriate. Here, the KMO is 0.772, therefore we conclude that a factor analysis is appropriate (Appendix C, table 1).

(36)

F.M. Hazelhoff – University of Groningen 36 extracted by the factors. According to Malhotra (2007) the criteria is an Eigenvalue of greater than 1.0 and the factors should extract at least a total of 60% of the variance. Factors with Eigenvalues of lower than 1.0 are not better than a single variable. As can be seen in Appendix C table 2, we therefore choose for four factors. Four factors have an Eigenvalue larger than 1 and cumulative variance of 68.60%.

Looking at the rotated component matrix (Appendix C, table 3) it can be seen that the items of perceived usefulness and relative advantage all have high factor loadings for component one. This indicates a correlation between these three variables. The results of a Pearson correlation analysis confirm the correlation between perceived usefulness and relative advantage (Appendix C, table 4). The correlation coefficient between perceived usefulness and relative advantage is 0.59. All the constructs are highly correlated with attitude. The other items of perceived ease of use, existing usage patterns and observability only has high loading on their own construct. We still choose to combine the items as it was originally meant by taking the average score of the items and transform the items into one variable. If we combined the items with high loadings on factor one into one variable we are not able to test the separate effect of the independent variable on the dependent variable as stated in the literature section.

In the second factor analysis, the items used for attitude, social norms, perceived behavioral control and trust in the online web vendor are checked whether those items can be put into one factor. The Bartlett’s test is significant (p = 0.000), rejecting the null hypothesis, and the KMO is 0.772, therefore we conclude that a factor analysis is appropriate (Appendix C, table 5).

Based on the Eigenvalues and the total variance explained we choose for four factors. The eigenvalue for four factors is just below one (0.997), however we still choose for four factors since 4 constructs are measured (Appendix C, table 6). The cumulative variance for four factors is above 60%, 76.55% as can be seen Appendix C table 6.

Table 7 in Appendix C gives an overview of the rotated component matrix. As can be seen in this table, the item attitude 3 has a loading of 0.655 with the first factor. The Pearson correlation analysis confirms the correlation between attitude and subjective norms (Appendix C, table 8). This might lead to multicollinearity between attitude and subjective norms, and will be tested in the regression analysis.

(37)

F.M. Hazelhoff – University of Groningen 37 reliability and factor analysis we can conclude that the items for attitude, subjective norms, perceived behavioral control and trust can be averaged and transformed into the four constructs.

4.3 Multiple regression analysis

As discussed in the previous section, three models are estimated by performing a multiple regression analysis. Before the model is estimated it will be tested if the model is linear and if non-normality, heteroscedasticity or multicollinearity is present. Autocorrelation is not tested, since that is only a potential problem for time-series models (Leeflang et al., 2000). If multicollinearity is present, it means that there is a relation between predictor variables (correlated), which can make it difficult to assess the relative importance of the independent variables in explaining the variation in the dependent variable (Malhotra, 2007). A violation of the other assumptions can lead to wrong estimates of the variance of the parameters which can result in wrong conclusion about the significance of the effect (Leeflang et al., 2000).

4.3.1 Assumptions

First, we will visually check whether the relationship between the dependent and independent variables is linear. If then relationship is non-linear, the results of the regression analysis will under estimate the true relationship. To test for linearity the residual plot of the standardized residuals and the standardized predicted values should be examined. The residual plots can be found in Appendix D, figure 1 and 2. Looking at these plots there are no non-linear shapes, meaning the linearity assumption is met.

Secondly, the models are tested for non-normal errors. It is assumed that the disturbance term is normally distributed. This can be tested with the Kolmogorov-Smirnov test and by calculating the Z-value of the skewness and kurtosis of the standardized residuals. If the Kolmogorov-Smirnov test shows non-significant p-values normality can be assumed. To calculate the Z-value of the skewness and kurtosis the statistic has to be divided by the standard error. For the skewness and kurtosis test a critical value of ±1.96 corresponding to a 5% significance level is used.

The breakthrough of Mobile commerce

The breakthrough of Mobile commerce

Examining the determinants of adoption intention of mobile commerce

The breakthrough of Mobile commerce

Examining the determinants of adoption intention of mobile commerce

Management summary

Preface

Table of contents

1. Introduction

2. Theoretical framework

3. Methodology

4. Results

Smartphone usage while shopping

Smartphone usage while shopping

Smartphone usage while shopping