The influence of personality on adoption of innovations : the case of Amazon Go

(1)

The influence of personality on adoption of

innovations: the case of Amazon Go

Student: Simone Ghislaine Kras Student number: 11398582

MSc Business Administration - Innovation and Entrepreneurship Track Amsterdam Business School, University of Amsterdam

Supervisor: dhr. dr. Tsvi Vinig Second supervisor:

(2)

1

Statement of originality

This document is written by student Simone Kras who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

2

Abstract

Different theories are explaining the decision-making process of people and there is a wide range of reasons for and against the decision to adopt a new product or service. This study contributes to a deeper understanding of the drivers preceding reasons for and against adoption. In this study, the role of personality as an antecedent for the norm and value barrier is investigated. A survey was conducted to measure Dutch consumers’ score on the Big Five inventory and norm and value barriers, as well as the willingness to adopt the innovation Amazon Go. Results show that none of the five personality traits influence the norm or value barrier, or the intention to use the service of Amazon Go. However, significant results are found for the relationship between the barriers and the intention to use. In line with previous literature, the results of his study suggest that these barriers negatively influence the intention to adopt.

(4)

3

1. Introduction

The time it took the internet to gain 50 million users was 7 years. The time it took the online interactive game Pokemon Go to gain the same number of users was 19 days. Other innovations will never even reach 50 million users. What are the reasons for people to resist or adopt a new product or service? Consumer adoption of new products and services is a subject which is researched multiple times: its outcomes provide valuable information for consumer business on how to bring their innovative product to the market. It may also predict how successful a new product will be.

At the start of this subject is the theory of reasoned action (TRA; Fishbein and Ajzen, 1975), that states that behavioral intention is the strongest predictor for actual behavior. Attitude toward that behavior and the perception of what other people might think of that behavior determine the intention. Years later formulated Davis (1989) another model, derived from TRA: technology acceptance model. This is a model that specifically focusses on the determinants of technological innovations. Two concepts were developed that are still used in innovation resistance literature: perceived usefulness and perceived ease of use of an innovation. The behavioral reasoning theory (BRT; Westaby, 2005), provides more insight in the reasons for or against performing a behavior and allows different psychological paths in decision making. Then there is another key theory that together with the previously discussed theories, form the theoretical background for this thesis: the diffusion of innovation theory (DOI; Rogers, 1962). This theory specifically assesses how the rate of innovation adoption is determined and formulates five perceived attributes of innovation: relative advantage, compatibility, complexity, triability and observability. Later I will discuss these key theories in detail.

(6)

5 Research discovered multiple reasons for people to resist or adopt an innovation. Some studies split the reasons into two groups: reasons for and against adoption. Other studies do not make this distinction. Ram and Sheth (1989) focus on barriers to innovation acceptance from a consumer perspective. Claudy et al. (2015) used their conceptual article to conduct a study on reasons for as well as against adoption of a sustainable innovation. One of their conclusions was that these reasons are not easy to generalize because they depend on the nature of the product of service. For instance, the strongest barrier to the intention to use a car-sharing service was different than the strongest barrier to the intention to use wind-turbines. Thus, we can conclude that there is a wide range of research on the reasons to adopt or resist innovation.

However, the drivers that precede reasons to resist or adopt innovation remain relatively uninvestigated. Previous studies have argued that a better insight in barriers to innovation adoption gives managers the opportunity to ‘differentiate their communication strategies’ (Antioco and Kleijnen, 2010). Research has proven that some personal traits influence resistance to general change (Oreg, 2003), and it would be interesting to look into the effect of personal traits on the intention to resist innovation. Until now, only the influence of personality traits on technology acceptance (Deveraj et al., 2008) has been studied, as well as innovativeness (Agarwal and Prasad, 1998) and regulatory focus (Herzenstein et al., 2007). Therefore, the effect of personality traits on barriers to innovation acceptance has not widely been investigated. How attitude toward adoption is formed is studied multiple times and the existing models like TAM and BRT provide valuable insights in this process. However, the role of individual traits preceding these attitudes could still deepen our understanding of this process. By means of a survey the influence of personality on two important barriers to innovation adoption of Dutch consumers is investigated.

This study contributes to the existing literature about consumer resistance to innovation. It provides a deeper understanding of the factors that influence the intention to accept or resist

(7)

6 innovation. More knowledge about potential consumers and their motivations to reject or adopt new products could identify relevant strategies to bring an innovation to the market and marketing strategies. It could also provide information about the potential success of a new product, and help companies decide on whether or not to fully develop a product and bring it to the market.

The study is structured as follows: first, I will discuss the acceptance models (i.e. TAM, TRA, BRT and DOI) described above to provide a theoretical background to the adoption of innovation. Then I will dive deeper into the reasons for and against innovation adoption to get an overview of the existing literature. Next, I will present the method for this study and collected data, after which the hypotheses are tested to examine the influence of personality traits on reasons to adopt an innovation. Subsequently, the results are being discussed. Finally, implications and limitations of this study are provided, as well as suggestions for future research.

2. Literature review

Rogers (1962) has described the innovation adoption process as “the process through which an individual or other decision-making unit passes from first knowledge of an innovation, to forming an attitude toward the innovation, to a decision to adopt or reject, to implementation of the new idea, and to confirmation of this decision.” There’s an abundance of literature on the likelihood of consumers to adopt innovations. The diffusion of innovation model of Rogers (1995), theory of reasoned action (Fishbein and Ajzen, 1975) and technology acceptance model (Davis et al., 1989) focused on reasons to adopt innovation. Other studies focus on factors that lead to resistance to innovation (Ram and Sheth, 1989; Kleijnen et al., 2009). Since the acceptance of innovations depend on the kind of product or service (Claudy et al., 2015), methods and analyses are often fit to the product or service that is being researched. The

(8)

7 literature does not agree on whether reasons for adoption are the opposite of reasons against adoption of an innovation. Davis et al. (1989), Wu and Wang (2005) and Agarwal and Prasad (1998) do not make a clear distinction between reasons for and against adoption. However, Claudy et al. (2015) argue that there’s ‘a growing body of evidence’ that reasons for and against adoption are not simply the opposite (Westaby 2005; Westaby and Fishbein 1996; Westaby et al., 2010). Chazidakis and Lee (2013) give an illustrative example of this: environmental advantage may be a reason for a person to adopt electric vehicles, but it is unlikely that harming the environment will motivate people to not adopt this innovation. Also, Kleijnen et al. (2009) do make a distinction by only investigating reasons against adoption of innovation. To provide a clear overview, I will first discuss the theories about innovation adoption. Then I will dive deeper into the specific reasons against innovation adoption, after which the literature about personality and its link to innovation adoption is synthesized.

2.1 Innovation adoption

Innovation studies that focus on the likelihood of consumers to adopt an innovation, mainly use the Theory of Reasoned Action (Fishbein and Ajzen, 1975) and Technology Acceptance Model (Davis et al., 1989) as a starting point. Both theories are rooted in the assumption that adopting or resisting an innovation is determined by negative or positive attitudes towards that new product or service, formed by the consumers’ evaluation of product/service attributes. I will discuss both of them below, after which definitions of reasons for adoptions are given, based on the diffusion of innovations theory (Rogers ,1971).

Theory of Reasoned Action

According to TRA (Fishbein and Ajzen, 1975), which is a theory used to predict and explain general human behavior, behavioral intention is the main predictor of performing behavior.

(9)

8 Behavioral intention is determined by two factors: attitude toward behavior and subjective norm. Attitude toward the behavior is a function of beliefs that performing the behavior leads a particular outcome and the evaluation of those beliefs (i.e. evaluation of the desirability of that outcome). Subjective norm is determined by a person’s normative beliefs that other people, such as family, friends or society, think that the person should or should not perform the behavior and the person’s motivation to comply with those other people. Fishbein and Ajzen later argued that other external factors might affect the established TRA relationships between attitudes, intentions, beliefs and behavior. Personality is treated as an exogenous variable, leading to beliefs related to behavior.

Technology Acceptance Model

Davis and colleagues (1989) proposed a model derived from TRA: TAM. In this study, they researched determinants of computer usage and adoption of new information technologies. This model is specifically developed for this purpose. According to TAM, the intention to adopt an innovation is strongly influenced by perceived usefulness of the innovation, while perceived ease of use has a smaller but still statistically significant direct effect on the behavioral intention. Perceived usefulness refers to the probability that using an innovation will increase a person’s performance. Perceived ease of use is defined as the degree to which a person expects the innovation to be free of effort. They found that eventually, TAM explained 51 percent of the total variance in behavioral intention, as opposed to 26 percent explained by TRA. Wu and Wang (2005) tested TAM and argue that perceived ease of use only indirectly influenced intention to use through perceived usefulness. This is in line with what Davis et al. (1989) found. They measured behavioral intention at two moments in time, and they found that one hour after introduction of the system, perceived ease of use had a direct effect. However,

(10)

9 after fourteen weeks, intention was directly affected by perceived usefulness alone. Perceived ease of use then only indirectly affected behavioral intention.

Devaraj et al. (2008) incorporated the Five Factor Model of personality into the context of TAM. They argued that personality would be related to specific beliefs about perceived usefulness of a particular technology. Subjective norms (TRA) are also included in their research, since the relationship between subjective norm and behavioral intention has received empirical support. They found evidence for the moderating role of conscientiousness on the relationship between subjective norms and intention to use technology.

Theory of Diffusion of Innovations

The theory of diffusion of innovation specifically assesses how the rate of innovation adoption is determined. ‘Diffusion is the process by which an innovation is communicated through certain channels over time among the members of a social system’ (Rogers, 2010: 1). According to the theory of the diffusion of innovations (DOI; Rogers, 1971), innovations have certain attributes that influence the spread (i.e. adoption) of an innovation. Although this is a theory about the rate of adoption, its definitions of perceived characteristics are used in research about innovation adoption (Joachim et al., 2017; Claudy et al., 2015). These five attributes are people’s perceptions of the characteristics of that innovation. First, relative advantage is the degree to which an innovation is perceived as better than the idea it’s supposed to replace. Second, the degree to which existing values, past experiences and needs of potential adopters are consistent with an innovation is known as compatibility. Making an innovation more compatible can be accomplished by naming an innovation and positioning it relative to previous ideas. Third, the perceived difficulty to use and understand an innovation is called

complexity. Fourth, trialability is the degree to which an innovation may be experimented with

(11)

10 people (Rogers, 2002; 2010). He argues that all these characteristics positively influence the adoption rate of an innovation, except for complexity, which has a negative influence. Thus, the higher the perceived complexity of an innovation, the lower the rate of an innovaton. Claudy et al. (2015) used these five factors in their study as ‘adoption factors’ and complemented them with the two factors based on TAM (Davis et al., 1989): perceived usefulness and perceived ease of use.

A meta-analysis was performed on the reasons to adopt innovations (Tornatzky and Klein, 1982). They followed Rogers’ (1971; 2002; 2010) definitions of these characteristics and concluded that only three innovation characteristics had a significant relationship to innovation adoption: compatibility, relative advantage and complexity (Tornatzky and Klein, 1982). They also confirmed that complexity was negatively related to adoption, which could thus possibly be seen as a barrier to innovation adoption. Agarwal and Prasad (1998) argue that perceived usefulness (TAM) is similar to relative advantage, and perceived ease of use (TAM) to complexity. Hence, three factors come forward as determinants to adopt an innovation, which I have summarized in table 1.

Table 1

Term Definition Citation

Compatibility

Degree to which existing values, past experiences and needs of potential adopters are consistent with an innovation

Rogers (1971; 2002)

Relative advantage (i.e. perceived usefulness)

Probability that using an innovation will increase a person’s performance

Agarwal and Prasad (1998); Davis et al. (1989)

Complexity (i.e. perceived ease of use)

Degree to which a person expects the innovation to be free of effort

Agarwal and Prasad (1998); Davis et al. (1989)

(12)

11 2.2 Reasons against innovation adoption

Behavioral intention models such as TAM and TRA have focused on belief concepts (e.g. behavioral, normative and control beliefs) to get more insight into factors influencing behavior but did not provide understanding of motivational mechanisms by addressing reason concepts (Westaby, 2005). According to the behavioral reasoning theory (BRT; Westaby, 2005), reasons serve as important linkages between beliefs, global motives, intentions and behavior. Global motives refer to attitudes, subjective norms and perceived control. An important assumption in this theoretical framework is that because reasons help individuals justify and defend their actions, they impact global motives and intentions. This promotes and protects an individual’s self-worth. Global motives are defined as broad substantive factors that consistently influence intentions across diverse behavioral domains. Attitude, subjective norm and perceived control are ‘estimated at a broader level of abstraction and have significant predicted intentions across numerous studies’ (Ajzen, 2001). Thus, global motives comprehend these three concepts. In contrast to global motives, specific behavior provides a specific context to beliefs and reasons for or against behavior. Like in TRA and TAM, intention is seen as the predictor for actual behavior. However, beliefs and values serve as important antecedents of reasons for and against behavior, as well as for global motives. Westaby (2005) argues that beliefs may have a direct influence on global motives because people may not always activate deeper reason justification mechanisms, but directly generate global motive perceptions. Intention is predicted by reasons for and against behavior and global motives. For example, a person may explain his or her behavior using context-specific reasons, regardless of his or her global motives toward that behavior (Westaby, 2005).

This theory of behavioral reasoning is important to keep in mind when assessing reasons against innovation adoption. It provides a complete framework of consumers’ decision making

(13)

12 by including context-specific reasons and in contrast to TRA and TAM, BRT does make a distinction between reasons for and against a specific behavior.

Reasons not to adopt an innovation could be called barriers. Literature regarding barriers to innovation adoption are summarized in table 2. Ram and Sheth (1989) wrote an influential article in which they introduced several barriers to innovation adoption from the consumer perspective. They split the barriers in two categories: functional and psychological barriers. This categorization is also used by Claudy et al. (2015) and Joachim et al. (2017). When consumers perceive significant changes from adopting an innovation, functional barriers more likely to occur. The following three barriers are functional barriers: Usage barrier refers to the degree to which an innovation is perceived as requiring changes in consumers’ routines. When an innovation is not compatible with existing practices or habits, it is likely that a consumer will resist the innovation (Ram and Sheth, 1989). Value barrier refers to the degree to which the value-to-price ratio of an innovation is perceived in relation to substitute products. There is no incentive for consumers to change their behavior unless the new product of service can provide a strong performance-to-price value (Ram and Sheth, 1989). The risk barrier is defined as the degree of uncertainty regarding economic, physical, functional and social consequences of using an innovation. Herzenstein et al. (2007) argues that risk is an extension of ‘traditional’ adoption frameworks like TRA or TAM. Adopting a new product or service involves uncertainty because there is a lack of information about the innovation. This lack of innovation and corresponding uncertainty also influences the perception of ease of use and usefulness of a new product or service. The literature agrees on four types of risk barriers, although under different labels. The higher the cost of an innovation, the higher the perceived financial risk. This type of risk is labeled economic risk by Kleijnen et al. (2009). Ram and Sheth (1989) mention new drugs as an innovation that may increase the physical risk barrier. Physical risk is not mentioned by Kleijnen et al. (2009) and Claudy et al. (2015) and is labeled

(14)

13 by Joachim et al. (2017) as personal risk barrier: perceiving an innovation as a threat to an individual’s physical condition or property. Functional risk refers to uncertainty about the performance of an innovation and social risk refers to the (possibly negative) reaction of other people on using the innovation (Ram and Sheth, 1989).

Psychological barriers arise through conflict with a persons’ prior beliefs. This second barrier category consists of tradition barrier and image barrier. The tradition barrier refers to the degree to which a consumer needs to deviate from established traditions to use the innovation. The greater the deviation, the greater the resistance (Ram and Sheth, 1989). This

Table 2

Term Definition Citation

Usage barrier

Evaluation that the innovation is not

compatible with existing workflows, practices or habits, and that requires changes in

customers’ routine.

Ram and Sheth (1989); Kleijnen et al. (2009) and Claudy et al. (2015)

Risk barrier The degree of uncertainty regarding

consequences of using an innovation.

Ram and Sheth (1989); Kleijnen et al. (2009) and Claudy et al. (2015) o Financial (i.e.

economic)

Uncertainty about monetary value of the innovation.

Ram and Sheth (1989); Kleijnen et al. (2009) o Performance

(i.e. functional)

Risk of innovation not functioning properly or reliably.

Ram and Sheth (1989); Kleijnen et al. (2009) o Physical (i.e.

personal)

Harm to person or property that may be inherent in the innovation.

Ram and Sheth (1989); Kleijnen et al. (2009)

o Social Risk of facing social ostracism or peer ridicule.

Ram and Sheth (1989); Kleijnen et al. (2009)

Value barrier Comparison of an innovation with its

precursor; a consumer thinks the new product does not produce a relative advantage.

Ram and Sheth (1989); Claudy et al. (2015); Moore and Benbasat, (1991)

Norm barrier (i.e. tradition barrier)

Evaluation that the innovation is conflicting with, for instance, family values, social norms or entrenched traditions

Laukkanen (2016); Ram and Sheth (1989); Kleijnen et al. (2009) and Claudy et al. (2015)

Image barrier (associations within mindset of consumer)

Negative association of an innovation with its origin.

Ram and Sheth (1989); Kleijnen et al. (2009) and Claudy et al. (2015)

Communicability barrier

Experienced difficulties in sharing an

innovation's benefits or shortcomings through language use

Moore and Benbasat (1991); Ram (1987)

(15)

14 barrier is labeled norm barrier by Laukkanen (2016) and defined similarly as ‘evaluation that the innovation is conflicting with, for instance, family values, social norms or entrenched traditions.’ The image barrier is defined as the degree to which an innovation is perceived as having an unfavorable image. The product class or industry to which an innovation belongs or the country where it was manufactured, provide the new product or innovation with a certain identity which may be favorable or unfavorable(Ram and Sheth, 1989).

Claudy and colleagues (2015) conducted two empirical studies on high-involvement product and service innovation, to examine whether the behavioral reasoning theory is a suitable framework to structure the mental processing of innovation adoption. Reasons for and against two specific innovations were studied: micro wind turbines and car sharing. Their survey was adapted to these two innovations through qualitative research and subsequently, they found different results for the two types of innovations. Doing this research, Claudy and colleagues focused on the same two types of barriers applied by Ram and Sheth (1989), split into five factors concerning resistance to innovation. Functional barriers are usage, value-to-price and risk barriers (in financial, functional and social sense). Psychological barriers refer to tradition and norm barriers, and image barriers. Their main conclusions are that a) reasons for or against adoption are context specific, concluded from the finding that the main barrier to turbines was costs, while the main barriers for car-sharing were safety and availability, and b) consumers follow different psychological paths when assessing the type of innovation. This is in line with the BRT, that assumes that people activate different psychological paths when making decisions in different (adoption) contexts (Westaby, 2005). The two studies are difficult to generalize since they consist of different samples and because the surveys specifically focused on the two innovations. Therefore, it is not possible to draw conclusions about the mediating role of consumer traits or contextual variables. This study points out the difficulty to generalize determinants of adoption. Another study done by Laukkanen (2016)

(16)

15 about barriers to adopt mobile banking, confirms this conclusion. He found that regarding mobile banking, the value barrier was most influential, which is different from the results found by Claudy et al. (2015). Although these determinants tried to encompass as many motivations as possible, the reason not to use mobile banking is fundamentally different from the reason not to use an innovative car-sharing service. It seems that for each kind of product or service, different determinants come forward as ‘most influential’. For instance, security issues and the corresponding physical risk barrier are more important in a car-sharing service than when using a new computer system.

Joachim et al. (2017) conducted a study on the antecedents of the resistance to product and mobile service innovations. They also found a difference between those two innovations, although in both cases, the norm barrier had the strongest influence on the intention to adopt. For product innovations, the value barrier was found the second strongest predictor for adoption. However, in the case of mobile service innovation adoption, the communicability barrier is the second strongest influencer. This barrier is defined as experienced difficulties in sharing an innovation's benefits or shortcomings through language use (Moore and Benbasat, 1991; Ram, 1987).

Kleijnen and colleagues (2009) focused on the effect of eight drivers on three types of resistance to innovation: postponement, rejection and opposition. The drivers are 1) traditions and norms, 2) existing usage patterns, 3) perceived image, 4) information overload, 5) physical risk, 6) economic risk, 7) functional risk and 8) social risk. They found that postponement of adopting an innovation is affected by economic risk and existing usage patterns. Rejection is affected by economic risk, existing usage patterns, functional risk, social risk and perceived image of innovation. Opposition to an innovation is affected by functional risk, social risk and perceived image. Physical risk and traditions and norms are reasons to oppose rather than reject. To study these forms of innovation resistance, a qualitative focus group study was

(17)

16 conducted. Quotes were collected and classified in one of the eight drivers. However, by only focusing on drivers of resistance, they assumed there was resistance to a particular innovation. This may have forced participants to come up with reasons not to adopt a product, while they may want to adopt this innovation in real life. This could have led to a bias in this research. This study also contributes to the BRT in the sense that consumers follow different psychological paths when forming their intention to behavior. Drivers of resistance of meaning differ per type of resistance and influence people’s decisions in different ways.

Hypotheses

The norm barrier is labeled tradition barrier or tradition and norm barrier, depending on the study. However, what it comprehends is always regarded to conflict with existing norms, values and habits. As described above, the predictive power of this barrier to intention to adopt an innovation is proven multiple times in the literature, although the salience of this barrier differed per innovation. The use of an innovation is inherently requiring individuals to accept changes in their habits and routines, and it may possibly put a strain on existing values and traditions. For instance, if it’s a tradition to collect wood for the fireplace on Christmas eve, and the innovation is an electric fireplace, it requires people to break with the tradition to collect wood. This may keep people from adopting such an innovation. Therefore, I hypothesize that for the intention to use Amazon Go:

H1: The higher the norm barrier, the lower the intention to adopt an innovation.

The value barrier refers to the relative advantage of the new product or service relative to that of other products or services. As described above, the predictive power of this barrier to intention to adopt an innovation is proven multiple times in the literature, although the salience of this barrier differed per innovation (Ram and Sheth, 1989; Claudy et al., 2015; Moore and

(18)

17 Benbasat, 1991; Laukkanen, 2016). For instance, the use of the innovative service mobile banking was best predicted by the value barrier (Laukkanen, 2016). Thus, I expect that for the service of Amazon Go:

H2: The higher the value barrier, the lower the intention to adopt an innovation.

2.3 Influence of individual traits

While there’s lots of literature on reasons to adopt innovations, less research is done on antecedents of these reasons. However, Claudy et al. (2015) did note the importance of contextual variables and consumer traits and suggested more research should be done on this subject. The BRT (Westaby, 2005) allows for the influence of individual traits as well, through beliefs and global motives. I will discuss some studies about the influence of individual traits such as regulatory focus, self-efficacy and personal innovativeness below, after which I will dive deeper in the literature about the Big Five personality traits and their link with reasons not to adopt an innovation. Subsequently, hypotheses will be formulated.

Herzenstein et al. (2007) studied the influence of a person’s regulatory focus on their decision making. This is defined as the motivational stage determined by a person’s socialization (mainly with caretakers) or by situational factors (for instance framing of task instructions). This regulatory focus can be either promotion-based or prevention-based and the authors argue that perceived new product risk does not influence all consumers in an equal way. The promotion system uses approach strategies to get a desirable outcome and is derived from nurturance needs. The presence of absence of positive outcomes is directive for promotion-focused consumers. In contrast to the promotion regulatory system, the prevention regulatory system is more sensitive to the presence and absence of negative outcomes. This system uses avoidance strategies to get a desirable outcome and is derived from security needs.

(19)

18 As one might expect, they find that prevention-focused consumers are more receptive to risk barriers regarding new products (Herzenstein et al., 2007).

Personal innovativeness is another individual trait of consumers that might influence

the intention to use a new product or service. Innovativeness can be defined in two ways. Global innovativeness, willingness to change, is a characteristic which has low predictive power for a specific innovation adoption decision (Goldsmith and Hofacker, 1991; Leonard-Barton and Deschamps, 1988). Agarwal and Prasad (1998) based their study on TAM and looked at the moderating role of personal innovativeness in the domain of information technology: the willingness of an individual to try out any new information technology. The inclusion of individual differences would provide insight into how perceptions are formed, that subsequently influence intention to use an innovation. They argue that only the characteristics relative advantage, complexity (perceived ease of use in TAM) and compatibility consistently relate to the intention to adopt an innovation. Agarwal and Prasad (1998) found that personal innovativeness only moderated the relationship between compatibility and the intention to use new information technology.

A direct relation between self-efficacy and resistance to technological change was found by a study done by Ellen et al. (1991). Self-efficacy is a judgment of one's own performance capability in specific settings. This subjective evaluation of competence or ability to perform the required task(s) or behavior is determined by the individual's interactions with and feedback from his/her environment and may not necessarily reflect actual competence or ability (Bandura, 1977, as cited in Ellen et al., 1991: 199). It has high predictive power when applied to behavior. They found that individuals who perceive low self-efficacy are more resistant to change.

Over the last decades, a valid five-factor model of personality emerged, which is also known as the Big Five. It groups specific personality traits in five basic personality dimensions:

(20)

19 Neuroticism, Extraversion, Openness to Experience, Conscientiousness and Agreeableness (Digman, 1990). I will briefly discuss them below and provide the hypotheses for this study.

Neuroticism

Neuroticism indicates a person’s emotional stability and ability to adjust. People who score high on this dimension experience relatively more anxiety, hostility, impulsiveness and vulnerability (as cited in Zhao and Seibert, 2006). They are less emotionally stable and expected to have less faith in their abilities to deal with change (Oreg, 2003), which relates to the findings of Ellen et al. (1991). Less faith in ability to change may indicate a low self-efficacy, which in turn is related to a high degree of resistance to innovations. Oreg (2003) found a positive correlation between neuroticism and resistance to change, because of the low faith in the ability to deal with change. People may feel threatened by the change (e.g. a new product or innovation and the corresponding required changes in behavior) and resist it. Thus, the expectation is that:

H3a: Neuroticism will negatively influence intention to adopt an innovation, mediated by the norm barrier.

H3b: Neuroticism will negatively influence intention to adopt an innovation, mediated by the value barrier.

Extraversion

Extraversion refers to the degree to which people are ‘assertive, dominant, energetic, active,

talkative and enthusiastic’ (as cited in Zhao and Seibert, 2006: 260). Oreg (2003) found a ‘low

yet significant’ negative correlation with routine seeking and short-term focus. Furthermore, extravert people are characterized as dynamic and sensation-seeking. This sensation-seeking trait may be a reason why they would have a higher intention to adopt an innovation (Oreg, 2003). Accordingly, I formulate the following hypotheses:

(21)

20 H4a: Extraversion will positively influence intention to adopt an innovation, mediated by the

norm barrier.

H4b: Extraversion will positively influence intention to adopt an innovation, mediated by the value barrier.

Openness to Experience

This trait characterizes people who are ‘intellectual curious and tends to seek new experiences

and explore novel ideas’. Someone who scores high on this factor, can also be described as

innovative (as cited in Zhao and Seibert, 2006: 261). Oreg (2003) found that individuals that are low on openness to experience, are more likely to score high on resistance to general change. These individuals are generally more risk-averse and less tolerant of ambiguity. This risk-averse focus can be related to the regulatory focus theory by Herzenstein et al. (2007), as described above: prevention-focused people are driven by security needs and are more receptive to risk barriers. Nov and Ye (2008) studied the influence of openness to experience on personal innovativeness in information technology and found a positive relation: the higher score on openness to experience, the higher the personal innovativeness in information technology. The explanation for this relation is the emphasis of this trait on creativity and non-conventional thinking. These findings suggest the following relations:

H5a: Openness to Experience will positively influence intention to adopt an innovation, mediated by the norm barrier.

H5b: Openness to Experience will positively influence intention to adopt an innovation, mediated by the value barrier.

Conscientiousness and agreeableness

Conscientiousness refers to the degree to which a person organizes, works hard, is persistent and motivated to accomplish a goal. A person’s interpersonal orientation is assessed in agreeableness. Trusting, forgiving, caring and gullible people score high on this factor. There

(22)

21 is no theoretical reason found in literature to expect that these two big five dimensions are correlated to intention to adopt Amazon Go. However, for the sake of completeness of this study, I intuitively hypothesize:

H6a: Conscientiousness will negatively influence intention to adopt an innovation, mediated by the norm barrier.

H6b: Conscientiousness will negatively influence intention to adopt an innovation, mediated by the value barrier.

H7a: Agreeableness will negatively influence intention to adopt an innovation, mediated by the norm barrier.

H7b: Agreeableness will negatively influence intention to adopt an innovation, mediated by the value barrier.

3. Data and method

The objective of this study is to examine the effect of personality on reasons for resisting an innovation. There is literature on resistance to innovation (Ram and Sheth, 1989; Kleijnen et al., 2009; Claudy et al., 2015, Joachim et al., 2017; Laukkanen, 2016) and research has been done on adoption of (technological) innovation (Davis et al., 1989; Rogers, 1971, Westaby, 2005). However, research on the drivers that precede reasons to resist or adopt innovation could still deepen our understanding of the subject. Deveraj and colleagues (2008) studied the antecedents of the technology acceptance model and found some links with personality. However, antecedents of reasons against innovation adoption remain uninvestigated. Claudy and colleagues (2015) conducted a research concluding that reasons for and reasons against adoption are not simply each other’s opposite. Also, they found that, in line with the behavioral reasoning theory (Westaby, 2005), reasons to adopt or resist an innovation are dependable on the product. Since knowledge of the decision to adopt or reject an innovation could influence perception of that innovation, it should be a predictive study (Tornatzky and Klein,1982).

(23)

22 Therefore, I will focus on the innovation Amazon go, which is a service that is not used in the Netherlands (yet). It is a grocery store where products a consumer picks are automatically added to the virtual cart. When the consumer walks out of the store, the new technology adds up the virtual cart and charges the consumers’ Amazon account. No checkout is needed.

To examine how each individual trait influences the innovation adoption, this study has the following research question:

What is the influence of personality on the intention to adopt an innovation, mediated by

norm and value barriers?

The conceptual model (see figure 1) consists of the dependent variable intention to adopt an

innovation and the independent variables neuroticism, extraversion, openness to experience,

conscientiousness and agreeableness based on the Big Five literature. The relative impact of

each trait is studied. I hypothesize that the relationship between the Big Five dimensions and intention to adopt is mediated by two reasons to resist an innovation. For feasibility reasons I selected two barriers that in the literature consistently have a strong influence on the intention to adopt or resist an innovative service: norm and value barrier (Joachim et al., 2017; Laukkanen, 2016; Claudy et al., 2015).

(24)

23 Method

For this cross-sectional study, I collected data by conducting a web-based survey. With this survey I gathered data about people’s personality, intention to use, and why they would potentially resist an innovation. The questionnaire was anonymous, which has the potential to decrease socially desirable responses. The surveys were administered digitally through social media and e-mail. A pilot-study was conducted to test the survey on flaws and did not lead to significant alterations.

Measurements

To measure personality, I will use the Big Five Inventory (BFI) 44-item scale. This scale is translated to Dutch and tested by Denissen and colleagues (2008). Cronbach’s alpha for each of the five personality traits is >.7. An example is ‘Ik zie mezelf als iemand die… Een werker

is waar men van op aan kan.’ (5-point Likert scale). For the sake of completeness, all five

dimensions were measured. For all items, see Appendix A. To measure the norm and value barriers, I have used 5-point Likert scales used by Joachim et al. (2017). The two scales consist of three items each. An example of a question regarding norm barrier is ‘Deze service past bij

mijn normen en warden.’ (r). An example of a question regarding value barrier is ‘Deze service lost een probleem op dat ik niet met vergelijkbare services kan oplossen.’ (r). The dependent

variable Intention to use is measured by a 5-point Likert scale based on the one created by Mackenzie et al., (1986), and consists of two items. An example question is ‘In de toekomst zie

ik mezelf de service gebruiken.’ (5-point Likert scale, α =.85). See Appendix B for all items on

the norm and value barriers and intention to use Amazon Go. The questions to measure the norm barrier, value barrier and intention to use needed to be translated to Dutch, for which the strategy of back-translation is used to ensure the quality of the questionnaire. I have recorded age, gender, education, and occupation as control variables commonly used in this research field (Claudy et al., 2015; Lian and Yen, 2013).

(25)

24 Sample

The population I have studied consists of Dutch residents. I have used a non-representative convenience sample because of the feasibility of this option. Respondents were selected by means of social media and e-mail. Demographic data (control variables) gathered from the respondents can be used to compare the sample to the population. In total 209 people have participated in this study, of which 174 fully completed the survey. The sample consists of 49 and 51 percent of men and women respectively. There is an overrepresentation of higher educated people: 74 percent has finished HBO/WO bachelor or higher. 24 percent of the respondents falls in the range of 18 to 24 years of age, and 29 percent in the range of 45 to 54 years. The other categories all represent around 15 percent of the respondents. Respondents below the age of 18 were not included in this study. 36 percent of the respondents was already

known with the innovation, while 62 percent did not have any knowledge about the innovation.

4. Results

Results of the gathered data are presented and discussed below. Cases with missing data were deleted listwise. The Big Five Inventory (Denissen et al., 2008) has multiple counter indicative items which have been recoded, so they now correspond with the rest of the items in terms of positive/negative responses. The measurements for the Norm barrier and Value barriers needed to be recoded as well, since a positive answer needs to correspond with a positive attitude towards the innovation, and therefore a low score on the barrier. Each personality traits consists of seven or eight indicators (Denissen et al., 2008). The mean of these indicators needed to be computed into new variables to get an overall score on each trait. This resulted in five new variables: Neuroticism, Openness to Experience, Agreeableness, Conscientiousness and Extraversion. These five represent the independent variables. The same was done for the indicators for the mediating variables Norm barrier and Value barrier and the dependent variable Intention to Adopt.

(26)

25 4.1 Data analysis

Table 3 shows the means, standard deviations of, and correlations between the variables. Norm barrier is negatively correlated with the intention to use Amazon Go, r = -.78 (p < .001). Value barrier is also significantly negatively correlated with the intention to use Amazon Go, r = -.67 (p < .001) and positively correlated with norm barrier, r = .70 (p < .001). The Big Five personality traits do not have a significant correlation with either intention to use, norm barrier or value barrier.

Table 3: Means, Standard Deviations, Correlations

Mean SD 1 2 3 4 5 6 7 8

1. Intention to use Amazon Go 3.52 1.03 (.946) 2. Norm Barrier 2.68 1.01 -.78** (.931) 3. Value Barrier 2.68 .79 -.67** .70** (.810) 4. Neuroticism 2.50 .60 -.01 .02 .12 (.828) 5. Extraversion 3.67 .60 .01 -.05 -.09 -.35** (.851) 6. Openness to Experience 3.53 .49 -.01 .00 .05 -.05 .25** (.743) 7. Conscientiousness 3.60 .48 -.10 .05 -.06 -.32** .15* .05 (.749) 8. Agreeableness 3.71 .47 -.09 .14 .02 -.38** .18* .18* .32** (.734)

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

To test the goodness of the scale of the norm barrier and value barrier, a principal axis factoring analysis (PAF) was conducted. The Kaiser–Meyer–Olkin measure verified the sampling adequacy for the analysis, KMO = .86. Bartlett’s test of sphericity χ² (15) = 787.990,

p < .001, indicated that correlations between items were sufficiently large for PAF. An initial

analysis was run to obtain eigenvalues for each component in the data. One component had an eigenvalue over Kaiser’s criterion of 1 and in combination explained 68.8 percent of the variance. However, examination of the scree plot revealed a levelling off after the second factor. Thus, two factors were retained and rotated with an Oblimin with Kaiser normalization rotation. Table 4 shows the factor loadings after rotation. The items that cluster on the same

(27)

26 factors suggest that factor 1 represents the norm barrier, and factor 2 the value barrier. As the results suggest, none of the items show high cross-leadings. Thus, no item had to be removed from the results.

Table 4: Principal axis factoring analysis

Item

Rotated Factor Loadings Norm Barrier Value Barrier

Deze service past bij mij .902 .053

Deze service past bij mijn normen en waarden .904 .031

Deze service past bij mijn persoonlijkheid .978 -.054 Deze service biedt voordelen die vergelijkbare

services niet bieden

-.119 .957

Naar mijn mening is deze service beter dan vergelijkbare services

.388 .584

Deze service lost een probleem op dat ik niet met vergelijkbare services kan oplossen

.105 .759

Eigenvalues 4.13 .75

% of variance 68.81 12.49

Rotation Method: Oblimin with Kaiser Normalization.

4.2 Testing the hypotheses

Having established the reliability of the variables, in a next step I will test the hypothesized relationships. To examine the mediating effect of the norm barrier and value barrier, I have followed the steps proposed by Baron and Kenny (1986), Judd and Kenny (1981), and James and Brett (1984). For each of the five personality traits I conducted an analysis to examine the relation between the independent variable (either neuroticism, extraversion, conscientiousness,

openness to experience or agreeableness) and the dependent variable (intention to adopt). This

first step is to assess whether there is an effect that may be mediated. The second step involves showing that the independent variable is correlated with the mediator (value barrier and norm

(28)

27 variable. Step 4 is to establish that the mediator completely mediates the relation between the independent variable and dependent variable.

Below, I will first examine H1 and H2: the relation between the mediators and the dependent variable. This assessment corresponds with step 3 (Baron and Kenny,1986; Judd and Kenny, 1981; James and Brett,1984). Then the rest of the hypotheses are tested by assessing step 1 and 2, aimed at studying the effect of the norm and value barriers acting as mediators. Together, these hypotheses contribute to answering the research question.

4.2.1 Norm and value barrier

Step 3: the effect of the norm and value barrier on the intention to use

To assess the relation between the norm barrier and value barrier and the intention to adopt, a hierarchical multiple regression analysis was conducted. The effect of the norm barrier and value barrier was controlled for age, gender and education. With this analysis the ability of the two variables to predict the dependent variable (intention to adopt) is tested (see table 5). In the first step of the hierarchical regression model, three control variables were entered: gender, age and education. The model is significant, F (3, 170) = 9.301, p < .05, and 14 percent of the variance in intention to adopt is explained. When the norm barrier and value barrier are added to the analysis in step 2, the total variance explained by the model as a whole is 63 percent F (5,168) = 58.048, after controlling for gender, age and education (R2 Change = .49; F (2,168),

p < .001). This means that the norm barrier and value barrier significantly explain an additional

49 percent in explained variance of intention to adopt. In the final model three out of five predictor variables were statistically significant. Norm barrier recorded a higher Beta value (β = -.58, p < .001) than value barrier (β = -.23, p < .001) and age (β = -.11, p < .05). In other words, if a person’s norm barrier increases with one, the intention to adopt will decrease with .58 points. Thus, H1 is supported. If a person’s value barrier increases with one, the intention

(29)

28 to adopt will decrease with .23 points. Thus, H2 is supported. And for every ten years increase in age, a person’s intention to adopt an innovation will decrease with .11 points. Gender and education were no statistically significant predictors of intention to adopt.

4.2.2 The effect of personality traits

I will discuss the results of H3 to H7 below. To examine if, and how, the norm and value barrier mediate the relation between the personality traits and intention to adopt, step 1 is to establish a relation between these independent and dependent variables. Step 2 is to show that the independent variable and mediating variables are correlated as well (Baron and Kenny, 1986; Judd and Kenny, 1981; James and Brett, 1984). Below, I will discuss the analyses conducted on the direct relationship between the independent and dependent variables (table 6), as well as the relationship between the independent variables and the mediating variables (table 7 and 8). Lastly, the consequences for the hypothesized mediating effect are discussed.

Table 5: Hierarchical Regression Model of Intention to adopt

R R2 _R2 Change B SE β t Step 1 .38 .14** Gender -.33 .15 -.16* -2.21 Age -.23 .05 -.34** -4.82 Education -.12 .08 -.11 -1.47 Step 2 .80 .63** .49** Gender -.06 .10 -.03 -.55 Age -.07 .03 -.11* -2.22 Education -.09 .06 -.08 -1.61 Norm Barrier -.58 .07 -.58** -8.59 Value Barrier -.30 .08 -.23** -3.55 Statistical significance: *p <.05; **p <.01 Dependent variable: Intention to adopt

(30)

29 Step 1: direct effect between the five personality traits and intention to use

I will discuss the Big Five dimensions one by one, based on the correlation analysis (table 3) and the hierarchical multiple regression analysis (table 6). The first step is to assess the correlation between the independent variable neuroticism and dependent variable intention to

adopt. There is no statistically significant correlation between neuroticism and intention to

adopt (r = -.01, p = n.s.). The regression analysis shows again that this trait has no significant effect on the intention to adopt (F (8, 165) = 3.442, p = n.s.). Secondly, no significant correlation between extraversion and intention to adopt was found in the analysis (r = .01, p = n.s.). Neither does the regression analysis show that extraversion has a significant effect on the intention to adopt (F (8, 165) = 3.442, p = n.s.). Thirdly, the correlation analysis shows no significant correlation between openness to experience and intention to adopt (r = -.01, p = n.s.). Neither does this personality trait contribute significantly to the total explained variance in intention to adopt (F (8, 165) = 3.442, p = n.s.). Fourthly, the same analyses were done for

conscientiousness, and there was no significant correlation was found between this trait and

intention to adopt (r = -.10, p = n.s.). Neither does the regression analysis show that this personality trait has a significant effect on the intention to adopt (F (8, 165) = 3.442, p = n.s.).

Table 6: Hierarchical Regression Model of Intention to adopt

R R2 _R2 Change B SE β t Step 1 .38 .14** Gender -.33 .15 -.16* -2.21 Age -.23 .05 -.34** -4.82 Education -.12 .08 -.11 -1.47 Step 2 .40 .16** .02 Gender -.28 .16 -.14 -1.74 Age -.25 .05 -.37** -4.90 Education -.13 .09 -.12 -1.49 Neuroticism -.23 .16 -.13 -1.45 Extraversion -.16 .14 -.10 -1.15 Openness to Experience -.01 .16 -.01 -.07 Conscientiousness -.08 .17 -.04 -.48 Agreeableness .06 .19 .03 .33 Statistical significance: *p <.05; **p <.01 Dependent variable: Intention to adopt

(31)

30 Finally, there is no statistically significant correlation between agreeableness and intention to adopt (r = -.09, p = n.s.). The regression analysis shows as well that agreeableness has no significant effect on the intention to adopt (F (8, 165) = 3.442, p = n.s.).

Step 2: effect between the five personality traits and norm and value barrier

Table 7 presents the results of the hierarchical multiple regression analysis of the independent variables and norm barrier and step 2 in establishing a mediating effect (Baron and Kenny, 1986; Judd and Kenny, 1981; James and Brett, 1984). The first model was statistically significant F (3, 170) = 8.73, p < .001 and explained 13 percent of variance in norm barrier. After entry of neuroticism, extraversion, openness to experience, conscientiousness and agreeableness in step 2 of the analysis, the total variance explained by the model as a whole was still 14 percent, F (8, 165) = 3.39, p < .05. Thus, introducing the five personality traits did not significantly contribute to the explained variance in norm barrier, after controlling for gender, age and education (R2 Change = .008, F (5, 190) = .91, p = n.s.). However, gender and age do show a significant effect on the explained variance in norm barrier. In the final model,

Table 7: Hierarchical Regression Model of Norm barrier

R R2 _R2 Change B SE β t Step 1 .37 .134** Gender .39 .15 .19* 2.57 Age .22 .05 .33** 4.56 Education .07 .08 .06 .79 Step 2 .38 .141** .008 Gender .34 .16 .17* 2.09 Age .23 .05 .33** 4.32 Education .09 .09 .08 1.03 Neuroticism .16 .16 .09 1.01 Extraversion .06 .14 .04 .42 Openness to Experience -.04 .16 -.02 -.25 Conscientiousness -.06 .17 -.03 -.35 Agreeableness .14 .19 .07 .75 Statistical significance: *p <.05; **p <.01 Dependent variable: Norm barrier

(32)

31 gender records a significant Beta value of .17 (β = .17, p < .05). Put differently, on average,

women scored .17 points higher on the norm barrier than men. And for every ten years increase in age, people score .33 higher on the norm barrier (β = .33, p < .001).

The same analysis was done for the value barrier, of which the results are shown in table 8. The first model was statistically significant F (3, 170) = 3.21, p < .05 and explained 5 percent of variance in value barrier. After entry of neuroticism, extraversion, openness to experience, conscientiousness and agreeableness in the second model of the analysis, the total variance explained by the model as a whole was 9 percent, F (8, 165) = 1.97, p < .05. Introducing the five personality traits did not lead to a significant increase in explained variance, after controlling for gender, age and education (R2_{Change = .03, F (5, 165), p = n.s).}

In the final model only one predictor of the value barrier was significant: age (β = .24, p < .05).

For every ten years increase in age, people score .24 higher on the value barrier. Other control variables were not significant predictors for the explained variance in value barrier.

Table 8: Hierarchical Regression Model of Value barrier

R R2 _R2 Change B SE β t Step 1 .23 .05* Gender .18 .12 .12 1.49 Age .11 .04 .20* 2.72 Education -.01 .07 -.01 -.18 Step 2 .29 .09* .04 Gender .16 .13 .11 1.27 Age .12 .04 .24* 3.01 Education .01 .07 .01 .08 Neuroticism .17 .13 .13 1.35 Extraversion -.01 .11 -.01 -.06 Openness to Experience .13 .13 .08 1.04 Conscientiousness -.12 .13 -.08 -.89 Agreeableness -.03 .15 -.02 -.18 Statistical significance: *p <.05; **p <.01 Dependent variable: Value barrier

(33)

32 The hypotheses

Step one and two of establishing a mediating effect (Baron and Kenny, 1986; Judd and Kenny, 1981; James and Brett, 1984) together give an answer to the proposed hypotheses H3 to H7. First, there is no significant correlation between neuroticism and intention to use and the analyses show that this trait does not significantly contribute to the explained variance of either intention to use, norm barrier or value barrier. There is no mediating effect taking place between neuroticism and intention to use. Therefore, H3a and H3b are both not supported. Secondly, no significant correlation was found between extraversion and intention to and this trait does not significantly contribute to the explained variance of either intention to use, norm barrier or value barrier. There is no mediating effect taking place between extraversion and intention to use. Therefore, H4a and H4b are both not supported. Thirdly, also openness to experience showed no correlation with intention to use. Furthermore, this dimension does not significantly contribute to the explained variance of either intention to use, norm barrier or value barrier. H5a and H5b are not supported since no mediating effect can be established. Fourthly, there is no direct effect between conscientiousness and intention to adopt, neither does the regression analysis show that this personality trait has a significant effect on the intention to adopt or the norm and value barrier. Thus, also with this trait no mediating effect can be established and therefore H6a and H6b are not supported. Finally, H7a and H7b are not supported either. There is no significant correlation between agreeableness and intention to use, and the regression analyses show that this trait is not a significant predictor for either the norm barrier and value barrier or intention to use.

Thus, to summarize, there is no direct relation between the independent and dependent variables. On top of this finding, the big five dimensions do not contribute significantly to the explained variance of either the norm or value barrier. Therefore, H3 to H7 are not supported.

(34)

33 However, in line with existing literature, H1 and H2 are once more confirmed. I will discuss the implications of these results in the next chapter.

5. Discussion

The aim of this study was to contribute to the existing knowledge about resistance to innovation. More specifically, this research strived to examine the mediating effect of two barriers on the relationship between personality traits and intention to adopt: norm and value barriers. With the assessment of this relationship, this study could contribute to the BRT, since this theory states that beliefs and values indirectly influence intention to behave through reasons for and against behavior and global motives (Westaby, 2005). Personality may or may not act as an antecedent for these reasons. Below I will apply the results of this study on the relevant literature and explain how the data relates to theory. Furthermore, I will formulate an answer to the research question and discuss the limitations of this study, along with suggestions for future research.

5.1 Discussion of findings

I will first discuss the findings regarding H1 and H2: the norm and value barrier. Next, I will discuss the results regarding H3 to H7: the five personality traits.

5.1.1 Norm and value barrier

How the intention of people to behave is determined is (partly) explained by multiple theories. The most general theory relevant to this study is the theory of reasoned action: attitude toward behavior and subjective norm are the main predictors of behavioral intention (Fishbein and Ajzen, 1975). The technology acceptance model is specifically focused on the consumers’ intention to adopt a technology. Perceived usefulness and perceived ease of use are argued to have a positive effect on intention to adopt a technological innovation (Davis et al., 1989).

(35)

34 Westaby (2005) argues that these theories are too simplistic for explaining consumers’ decision-making process. In the behavioral reasoning theory, Westaby (2005) incorporates more variables into the model. In BRT, behavioral intention is determined by global motives (which comprehends attitude to behavior, subjective norm and perceived control) and reasons for and against behavior. Global motives are determined by reasons and beliefs and values, and reasons are determined by beliefs and values. This theory allows different psychological paths in the behavioral decision-making process since people can ‘skip’ assessing their global motives in forming their behavioral intention for instance. This is confirmed by the research done by Claudy et al. (2015).

The results of this study confirm the relationship between the reasons against behavior and behavioral intention. The regression analysis showed a significant negative relation between the norm barrier and the intention to adopt, and the value barrier and the intention to adopt. The norm barrier (i.e. tradition barrier) refers to the degree to which a consumer needs to deviate from established traditions to use the innovation. The greater the deviation, the greater the resistance (Ram and Sheth, 1989). This barrier emphasizes the family values, social norms or entrenched traditions (Laukkanen, 2016). In this study, the norm barrier turned out to be the strongest predictor of intention to adopt Amazon Go. The higher the score on the barrier, the lower the intention to adopt. In other words, the more a person feels that Amazon Go fits his or her personality and norms and values (i.e. a low score on the barrier), the more likely he or she would adopt this innovation (i.e. a high score on intention to use). This is in line with previous research. Joachim et al. (2017) found a statistically significant negative relationship between norm barrier and intention to adopt. This result was found for product innovations as well as mobile service innovations and turned out to be the strongest predictor for intention to adopt. Laukkanen (2016) found that the tradition barrier (i.e. norm barrier) leads to rejection of internet banking.

(36)

35 Besides the impact of the norm barrier, I also found a statistically significant relationship between the value barrier and intention to adopt Amazon Go. The value barrier refers to the relative advantage of the new product or service relative to that of other products or services (Ram and Sheth, 1989). It is also a predictor for the intention to adopt this service. Similar to the results regarding the norm barrier, the higher the value barrier, the lower the intention to adopt. Put differently: the more a person perceived Amazon Go as better than other similar services (i.e. a low score on the value barrier), the more likely he or she is to adopt this innovation (i.e. a high score on intention to use). This result is in line with previous research on other innovations. Laukkanen (2016) found that the value barrier is the greatest impediment to mobile banking adoption. Joachim et al. (2017) also found a significant negative effect of value barriers on the intention to adopt product innovations and mobile service innovations.

Age negatively influences the probability of a person of the intention to adopt Amazon Go as well. In other words: the older a person, the lower the intention to adopt Amazon Go. A similar result was also significant in the research done by Laukkanen (2017), who also found that age had a negative effect on intention to use mobile banking. The fact that in this study no significant effect was found for education could be a result of an overrepresentation of higher educated people. 72 percent of the sample has finished HBO bachelors or higher, which is not representative for the population.

5.1.2 Personality traits

Claudy et al. (2015) suggested to investigate the relation between personality, reasoning, and intention to adopt innovations. Previous literature provided information about a possible effect of personality traits on the intention to adopt innovations. I hypothesized that this relationship was mediated by the norm and value barrier, which I have discussed before. However, the results of the data analyses show that there is no statistically significant relationship between

(37)

36 any of the personality traits and the intention to adopt Amazon Go. No significant correlation was found between the traits and intention to adopt. Neither did they contribute significantly to the total explained variance in intention to adopt an innovation. In other words, being extravert or introvert for instance, does not significantly influence a person’s intention to use Amazon Go. Differences in personality also do not influence a person’s score on the norm or value barrier. Therefore, both the value and norm barrier do not have a mediating effect between the traits and intention to adopt. Interestingly, this is not in line with what previous literature suggested. Oreg (2003) did find a significant relationship between neuroticism and resistance to change: more neurotic people were more likely to resist general change. On the other hand, he found a negative relationship between extraversion and openness to experience and resistance to change.

A possible explanation for this result may lie in the notion by Claudy et al. (2015) that the reasons that lead to the intention to adopt innovations are context-specific. This implies that for other innovations the relationship between personality and intention to use could be significant. Thus, being extravert (or neurotic, or open to experience, etc.) may influence the intention to use other innovations. For example, Devaraj et al., (2008) found a significant relationship between agreeableness and perceived usefulness of technology. People who are considered agreeable, are focused on cooperation with others and task accomplishment. Thus, it makes sense that these people are more likely to have the intention to adopt innovative technology that fosters cooperation and collaboration. However, for innovations that do not foster this, it makes sense that this trait does not influence the intention to adopt. Furthermore, the sample of the cross-sectional survey is not representative for the population. This may be another explanation for the absence of a statistically significant effect. This is a general limitation of this study, which I will explain in more detail below.

The influence of personality on adoption of innovations : the case of Amazon Go