1
-
Master’s Thesis
A STUDY ON THE TECHNOLOGY ACCEPTANCE OF VIRTUAL REALITY THROUGH THE HMSAM MODELGellert Kovacs – 11388447
Supervisor: Prof. Peter Van Baalen
Course: MSc Business Administration / Digital Business Date of Submission: 23/06/2017
2
Abstract
The objective of this study was to evaluate what factors influence consumers’ willingness and intention to adopt Virtual Reality devices into their daily lives. Technology adoption models have been researched and evaluated in great detail over the years, however specific research applying them to modern technological advancements such as Virtual Reality (VR) are still limited. VR introduces new obstacles and challenges for adoption into daily use, and old models might be insufficient in explaining these. A demonstration of VR was held among a total of 105 students and working professionals to participate in, and a questionnaire administered to measure participant’s perceptions of the experience. The results from the survey were analyzed through SPSS 22.0 through explanatory factor regression analysis, as well as testing moderation and mediation via the ‘PROCESS’ add-in. Of the 8 hypotheses, 3 proved to be significant: Perceived Ease of Use, Joy and Health Issues. Of the other factors, Perceived Usefulness could not be evaluated due to it low scale reliability, while Curiosity proved to be statistically non-significant, an outcome that is in contrast to other studies and models such as the hedonic motivation (HMSAM) model. Overall, the findings showed that certain aspects of older technology acceptance models (such as Perceived ease of use and Joy) are still applicable and useful in explaining adoption intentions today, while others (Curiosity) are not. In addition, evidence was found of the relevance of Health Issues as a moderating variable, and its inclusion in future technology acceptance models is suggested as a final amendment.
3
CONTENTS
Introduction ... 5
What is Virtual Reality... 6
State of Virtual Reality ... 9
Literature Review ... 10
Do people want VR? ... 10
Innovation Adoption ... 12
Diffusion of Innovations ... 12
Technology Acceptance Model (TAM) ... 13
Universal Theory of Acceptance and Use of Technology (UTAUT) ... 14
UTAUT (2) ... 14
HMS and HMSAM (Hedonic Motivation System Adoption Model). ... 15
Conceptual Framework ... 17
Research goal ... 17
Research Question ... 17
Theory Development (Hypotheses): ... 18
Method ... 21 Sample ... 21 Measure ... 23 Procedure ... 25 Results ... 27 Descriptive Statistics ... 27 Correlation Analysis ... 28
Collinearity VIF Analysis ... 29
Reliability tests ... 30
Regression Analysis ... 31
Moderation ... 35
Discussion... 36
Limitations & Future Research ... 40
Conclusion ... 41
Bibliography ... 42
5
1.0 Introduction
Virtual Reality (VR) is not a new phenomenon. The concept has existed for decades, and simple versions of consumer products have been available in the past. However, earlier consumer attempts with the technology have never been able to achieve a level of quality that delivers a true sense of engagement and presence. Virtual reality aims to transport viewers into a computer rendered environment using a headset, often allowing them to look around and interact with that virtual world as though they were a part of it. Oculus VR’s ‘Rift’ is the first product in recent times that promises to deliver such an experience. Oculus claims great VR as being able to achieve a sense of ‘presence’ within a virtual environment.
2014 was a crucial year for the company Oculus VR. Within 3 years, they grew from a college dropout project into a popular startup technology company that had inspired people around the world. The initial Kickstarter campaign for their first set of virtual reality goggles exceeded its $250,000 goal within a few days and went on to raise $2.2 million in total (Kickstarter.com, 2012). Following important hires and the successful delivery of the promised prototype developer kit (DK 1 and DK 2) units, Oculus VR went on to raise a further $75 million in seed funding to prepare for the launch of its first consumer version model (CV1).
However, virtual reality is a highly complex system to piece together, requiring hundreds of millions of dollars in funding and thousands of engineering hours to develop (TomsHardware, 2016). As far as consumer adoption of the technology is concerned, there is no concrete proof that a critical mass of people will be interested in investing their time and money into the use of VR. Companies are betting that people will, but they face a difficult task in convincing the public that VR isn’t just the next ‘fad’ that 3D TVs were. One of the tools available to companies to measure public interest is to examine technology adoption models. These models help explain what characteristics of new technologies foster their adoption and ultimate success. Multiple models exist; however, many of these were developed during a time when the technology landscape was vastly different. Additions and alterations to these models have been made in the past, and this paper will examine various models to gauge their relevance to VR today. A further
6 inspection of the history of 3D TVs can serves as a relevant example, as that technology was also developed in recent years and featured new technology centered around media consumption. Many studies exist that examine technology acceptance models, however due to the novelty of virtual reality, there is very limited literature regarding their acceptance and adoption. With the aid of a live demonstration and consequent questionnaires, this paper will add to the existing literature on technology acceptance models by examining which characteristics of these models are still important for VR. Furthermore, it will test additional criteria that are potentially important to VR, as well as propose any findings as supplementary characteristics to existing technology acceptance models.
1.1 What is Virtual Reality
Virtual Reality is the culmination of years of various technologies coming together. In its most basic form, it is a set of goggles placed on a person’s head with a digital screen incorporated within. By using a special ‘Fresnel lens,’ the viewer is able to focus his/her eyes on the close screen, which is made to appear as though it were much further away. Ideally, this screen is wide enough to fill a person’s field of view, in essence replacing the user’s visible world with a digital one. A special gyroscope tracks the head-movement of the viewer and updates the image inside the headset to reflect that head-movement. Unlike traditional TVs and monitors that are stationary, this enables the user to ‘look around’ within a virtual space by moving their head. However crude, basic versions of this technology have existed for many years. The reason VR has recently regained its popularity is due to critical advancements made in the smartphone industry. The LCD panels found in modern smartphones have become larger, brighter, faster, higher resolution, and most of all, cheaper. Gyroscopes have also become smaller, lighter, more accurate and cheaper. Together, these are essential traits without any of which modern VR would not be possible. When working in unison, they are able to eliminate most of the ‘deal-breakers’ that VR faces: motion sickness, eyestrain and brain strain. However, even according to Gabe Newell (2017) (founder of Valve and the HTC VR headset), “current iterations of VR products only just barely meet the minimum viable standards of
7 useable quality; for VR to become truly enjoyable, many more technology hurdles will have to be overcome.” Another crucial limitation that has been tackled by modern VR devices is new user input interfaces. Previously, users were limited to being passive observers of the virtual worlds. With Oculus’s new external tracker and ‘nun chuck’ controllers, head and hand tracking with full six degrees of freedom (6DoF) have also been implemented. This means a person can be fully tracked, and they are able to walk around in a virtual space and interact with virtual objects using their hands. The downside of this is that it also further complicates the physical assembly of the device. See image below of basic components:
The final element beyond the VR hardware is the actual virtual world. It is not enough to have the physical goggles; a computing device is also required to create the virtual world. Computer hardware and graphics processors have become increasingly powerful; to the point where rendering near-photorealistic scenes in real-time on traditional HD monitors has started to become accessible even to average consumers. However, “VR hardware is incredibly graphically taxing, requiring every single ounce of computational power available … from the most advanced, enthusiast class computers.” (Newell & Valve Employees, 2017).
Figure 1: Essential components of the HTC Vice VR system. Each item must be set up, placed, calibrated and synchronized to work in unison with all other components of the VR assembly (The Post5, 2016)
8 The specifications for a VR-Ready
computer have been set at the highest enthusiast level, with an entry price for the minimum computer starting at €1.299 and ranging to over €2.000+ (not including the €800 cost of the Vive or Rift VR headset) (Neiger, 2016). It is only once all of these components (VR headset, controllers,
trackers, computer etc.) are synchronized and calibrated that any experience may begin. In its current state, this setup is not a simple process and requires extensive technical know-how. As such, the experience is limited to early adopters and enthusiasts in its current form.
Once the unit has been set-up, the experiences the viewer can engage in are highly varied. From taking diving trips among schools of fish in the pacific to walking on Mars; virtual reality offers an ample selection of locales and activities to participate in. The Climb – one of the stand-out applications of the platform - allows users to experience the thrill of rock-climbing the cliffs and overhangs of lush tropical mountains, all from the safety of their living room. Users can engage in competitive sports over the internet, with a very high level of realism. Further potential applications include medical training simulations, architectural visualizations, virtual socializing and many others. Google is already offering education centers virtual, class field-trips to over 200+ locations (Rhodes, 2016). A school in Reston, Va. recently participated in a VR expedition to the LHC (Large Hadron Collider) particle accelerator at CERN in Switzerland. Although there are still many issues facing VR, experts agree that the technical challenges
Figure 3: (The Climb for Oculus Rift, 2016)
9 will be overcome within the next few years (Newell & Valve Employees, 2017). The real challenge lies in the public’s interest and willingness to adopt VR.
1.2 State of Virtual Reality
If successful, Virtual Reality (VR) has the potential to impact a wide array of applications, more so than many other technologies in recent years. The gold standard of new technologies has – in the past - been dictated by whichever medium the adult entertainment industry (AEI) adopted. This was how Blu-ray, DVD, VCD, VHS and many other digital standards before it won their, even though there were arguably better competing standards in existence. Due to the nature of VR, the user is exposed to experiences that are beyond mere passive observation. Interaction is key. In theory, this suits the AEI, however it requires high-end, powerful (and expensive) computational infrastructure. As such, the video game sector is initially better suited to adapt VR technology. As of 2014, the video game industry had a market capitalization of $21 billion (Entertainment Software Association, 2015) with a large consumer base that has the capacity to attract large corporations and warrant them to fund intense R&D projects. This is precisely the opportunity that Facebook also recognized.
In 2015, Facebook acquired Oculus for $2 billion, betting on VR to be the future of communication and entertainment. Not only did this provide Oculus with the capital to continue work on their Rift prototype, but it also legitimized VR in the eyes of the market and placed a monetary value on the technology. In the years since Facebook’s acquisition of Oculus, other companies have also announced and shipped various competing VR headsets. Valve has partnered with HTC to release the Vive (often considered the best VR experience currently available), while Sony has also released their version of the PlayStation VR. Microsoft and Intel each have products under development, along with countless other electronics manufacturers. It is clear that large corporations are betting on the success of VR. However, a similar trend could be observed during the initial development and release of home 3D TV. Now, eight years after the first commercial availability of a 3D Full HD Home Theater, Samsung and LG have
10 announced at CES 2016 (Consumer Electronics Show in Las Vegas) the discontinuation of 3D support on their premium LCD TV lines (Katzmaier, 2016):
“3D, once hailed as a breakthrough new feature on TVs and propelled into mainstream consciousness by the blue aliens of "Avatar" and the efforts of ESPN and DirecTV, has been
waning in popularity for years. Now it has absorbed that most telling of deathblows from the biggest gun in the TV hardware business. With a bullet to the head from Samsung, 3D
TV is now deader than ever”
VR is still in its infancy, to avoid such an outcome for Virtual Reality, it is important for proponents of the technology to know what went wrong with home 3D TVs and why consumers failed to embrace them. Innovation diffusion theories and technology acceptance models (TAM) could be the key to answering these questions. However, some of these models were first drafted in 1983 which – considering the rate of technological advancement – could mean they are well outdated. Consumer-level Virtual Reality is still an open field – one that has not been tried and tested yet. With such investments pouring in from the likes of Facebook, it would be imperative to know what factors influence consumers’ willingness to adopt Virtual Reality into their daily lives.
2.0 Literature Review
2.1 Do people want VR?
There is obvious interest for VR – that much is clear from the overwhelming success of the initial Kickstarter campaign (Eadicicco , 2014). However, exactly how large the market is for such a device will remain unknown until real consumer grade units are delivered. It is clear from reading technology websites that opinions about VR are wildly different. There are individuals who passionately patronize virtual reality as the future of digital entertainment. On the other hand, there also exists a severe distrust and apprehension among some people related to VR. A comparison is often drawn between VR goggles and 3D TVs, labelling both as a gimmick or a passing ‘fad.’ Concerns have been voiced over the need to wear
11 “clunky goggles,” with some people going so far as to “vow never to wear them.” The truth most likely lies somewhere between these extremes.
The trouble with VR is that the experience is extremely hard to describe with just words; Carmack (2014) (cofounder of id Software – current CTO of Oculus) has said that “virtual reality has to be experienced to be understood.” The sense of ‘presence’ in a virtual world is something that cannot be explained in an article or a video review (Gaudiosi, 2014). Large digital entertainment news websites are echoing these sentiments, such as Gamespot.com’s VR-reluctant editors changing their stance and stating that “Virtual Reality is the Future;”
“Little did I know that I was walking into a demo that would forever alter my opinion of VR and its potential to change not only gaming, but our relationship with technology at large. A minute into the demo, my reluctance transformed into excitement, and I was instantly converted into a VR believer. (Walton & Brown, 2014)
Statistics show that interest in VR is highly dependent on a person having tried it: According to Touchstone Research (2016) – a technology research group - interest in VR is highly dependent on a person having tried VR. Nearly 50% of those who have, said they are ‘impressed’ or ‘extremely impressed,’ with only 17% ‘not/not at all impressed’ (the rest identified as neutral). However, the statistics also show that consumer awareness is still relatively low. Of the group of 7021 people surveyed by Touchstone Research, only 41% stated that they were even aware of VR, and only 8% of those had tried it. (Drew, 2016). Even among enthusiasts, this number only stands at ~25.5% (Tom's Hardware, 2016). Of those that have experienced some form of modern VR, 57.1% changed their perspective on the issue of costly VR hardware.
12
2.2 Innovation Adoption
Modern technological innovations are rapidly changing the way we live our lives. However, Bala & Venkatesh (2013) assert that innovation for the sake of innovation is useless unless these new technologies are embraced by real people and are actively implemented into their ever-day routines. Throughout the last decade, many theories and models of innovation adoption have been conjured to predict and explain users’ adoption of new technology. One such example is the famous Technology Acceptance Model (TAM) (Davis, 1989) that states that beliefs and attitudes toward information systems are a driving force behind consumers’ willingness to adopt new technology. Other models include the Innovation Diffusion Theory (IDT) by Rogers (1995) and the Theory of Planned Behavior (TPB) (Mathieson, 1991).
2.2.1 Diffusion of Innovations
The unique traits of individual innovations also naturally play a large role in consumers’ willingness to adopt a new technology or not. These qualities were summarized by Rogers (2003) into 5 characteristics (as perceived by individuals) to help explain the differences in the adoption rates for various innovations:
Relative Advantage: the “degree to which an innovation is perceived as being better than the idea it supersedes (p. 15).
Compatibility: the “degree to which an innovation is perceived as consistent with the existing values, past experiences and needs of potential adopters” (p. 15).
Complexity: the “degree to which an innovation is perceived as relatively difficult to understand and use” (p. 15).
Trialability: “the degree to which an innovation may be experimented with on a limited basis” (p. 16)
13 Rogers (2003) argues that innovations that are perceived to have a higher relative advantage, compatibility, trialability, observability while being less complex will be more readily adopted by consumers. What this means for Virtual Reality is that in order to overcome the substantial apprehension of the general public about the technology, these goals will have to be improved and better communicated and marketed.
Of the above mentioned criteria, complexity and trialability are immediately the most apparent obstacles to the adoption of VR. Being the new tech that it is, setting up and running a VR system is complex, time consuming and expensive. This naturally leads to difficulties with trialability, something which is crucial for consumers to see the value in VR as discussed by John Carmack.
2.2.2 Technology Acceptance Model (TAM)
While Rogers’ (1995) TPB model deals with consumers’ perceptions about the characteristics of an innovation, other models such as TAM by Davis (1989) focus more on user intentions influencing their decision to adopt a new technology. These are most notably:
Perceived usefulness (PU) – "the degree to which a person believes that using a particular system would enhance his or her job performance".
Perceived ease-of-use (PEOU) – "the degree to which a person believes that using a particular system would be free from effort"
14
2.2.3 Universal Theory of Acceptance and Use of Technology (UTAUT)
Based on the literature above, it is probable that old models/theories are not suitable to explain consumers’ adoption habits of modern technology anymore (or they should be adapted). It was specifically for this reason that Venkatesh et al. (2003) formulated the universal theory of acceptance and use of technology model (UTAUT), incorporating eight theories on technology adoption; four main constructs – performance expectancy (PE), effort expectancy (EE), social influence (SI) and facilitating conditions (FC) – and four moderating variables; gender, age, experience and voluntariness of use.
2.2.4 UTAUT (2)
This was later further enhanced and updated to become the UTAUT2; the latest iteration of this
model. It differs from UTAUT in that it makes the distinction to include hedonic motivation systems (HMS)
as well as habits and price value as explanatory variables. Regarding virtual reality, the hedonic motivation aspect of the model is crucial in explaining a portion of peoples’ desire to engage with the device for reasons other than productivity.
Figure 5: Four main constructs of UTAUT (2) (Venkatesh et al. (2013)
15
2.2.5 HMS and HMSAM (Hedonic Motivation System Adoption Model).
Hedonic Motivation Systems (HMS) refers to “the influence that pain or pleasure has on a person’s
willingness to move towards a goal or away from a threat” (Lowry, Gaskin, Twyman, Hammer, & Roberts, 2012). The HMS model (figure 4 below) was first devised by Van der Heijden (2004) as a way to explain
hedonically motivated adoption characteristics. It made use of Perceived Usefulness (PU) and Joy as
variables to mediate the relationship between Perceived Ease of Use (PEOU) and Behavioral Intention to
use (BIU). Van der Heijden (2004) established that “...for utilitarian systems, we can expect extrinsic
motivation to be the dominant predictor of intentions to use the system—at the expense of intrinsic motivation. Similarly, for hedonic systems, we can expect intrinsic motivation to be the dominant predictor of intentions to use the system—at the expense of extrinsic motivation.”
This HMS model was later adapted specifically to technology by Lowry et al. (2012) to form the Hedonic
Motivation System Adoption Model (HMSAM). It provides not only a general extension to TAM, but a
HMS-specific acceptance model employing cognitive absorption (CA) as a key mediator of PEOU and
Figure 6: Original HMS Model (Lowry, Gaskin, Twyman, Hammer, & Roberts, 2012)
16
BIU. Lowry et al. (2012) breaks cognitive absorption up into five constituent elements: Control, Curiosity,
Joy, Focused immersion, Temporal dissociation, as shown below.
In contrast to Utilitarian Motivation Systems (UMS) that deal with satisfying basic needs that are
concerned with achieving specific outcomes of use rather than the process itself, “HMS—such as video games, social networking sites, and virtual worlds—can create a level of deep immersion and devotion
rarely seen with UMS” (Lowry, Gaskin, Twyman, Hammer, & Roberts, 2012; Jegers, 2007; Sherry, 2004). “Additionally, users who devote time to HMS do so for intrinsic rewards and generally have little concern for any potential external reward they might receive (Sweetser & Wyeth, 2005); rather, they are concerned
Figure 8. Overview of HMSAM, from Lowry et al. (2013)
Figure 7: Expanded HMS model by Lowry et al (2012)
17 mostly with the process or experience of use itself” (Lowry, Gaskin, Twyman, Hammer, & Roberts, 2012). After conducting their study, Lowry et al. (2012) arrived at the final model as shown in figure 8 above.
3.0 Conceptual Framework
While many theories and models have been developed to explain and predict consumers’ innovation adoption attitudes, most of these are relatively old theories in the context of modern technological advancements. Updates to the old models in the form of UTAUT and UTAUT2 have been made, however with the advent and proliferation of the always connected smartphone and Internet of Things (IoT) devices, technology and consumers’ user habits have so vastly changed over the last decade that it calls into question the applicability of these technology and innovation acceptance models. Virtual Reality further complicates matters, and due to its novelty, research relating to the adoption and diffusion of VR is extremely limited. This paper will attempt to answer what factors determine users’ willingness and desire to adopt Virtual Reality into their daily lives with a focus on the HMSAM model; as it already includes old and new adoption factors and is the most recent and relevant model for Virtual Reality. Therefore, the research goal of this paper is to:
3.1 Research goal
Investigate which factors of existing technology acceptance models (specifically the HMSAM - Hedonic Motivation System Adoption Model) are still applicable in explaining modern innovation adoption trends today with regards to new technologies such as Virtual Reality; and whether it needs to be amended to become useful again.
3.2 Research Question
Investigate which factors of the HMSAM model influence consumers’ willingness and intention to adopt Virtual Reality devices into their daily lives.
18
3.3 Theory Development (Hypotheses):
As discussed by (Van der Heijden, 2004), perceived usefulness is a construct that is dependent on the context in which it is being assessed: in utilitarian applications, perceived usefulness is judged based on the external goals that a system aids its user in achieving. These are goals external to the model, such as achieving increased productivity in a work environment through the use of a system. In a hedonic context however, perceived usefulness deals with goals that are internal to the user: such as experiencing joy and happiness through engagement with a system. In situations with low ease of use, subjects might become frustrated with their interaction with a system, thus lowering their perceived usefulness of that system. Indeed, Atkinson Kydd (1997), and Moon and Kim (2001) all found significant positive effects of ease of use on user engagement with various hedonic systems, suggesting a link between perceived ease of use and perceived usefulness. Thus, it is hypothesized that:
H1: There is a positive relationship between PEOU and Perceived Usefulness.
Motivational theory indicates that individuals engage with systems for one of two reasons: to obtain an outside benefit (extrinsic motivation) - or to gain a benefit that is internal to the user-system interaction (intrinsic motivation). “The definition of perceived usefulness draws attention to an outside benefit, external to the system-user interaction” (Van der Heijden, 2004) such as being able to ‘think more clearly by using the system.’ In such cases of extrinsic motivation, it logically follows that a system which is useful and aids its user in achieving desired external outcomes will motivate that user to revisit the system with the intention of using it again in the future. Thus, it is hypothesized that:
H2: There is a positive relationship between Perceived Usefulness and BIU.
Published, publicly available research is lacking in the area of health issues (nausea, dizziness, headache) stemming from VR use - or even just motion controlled electronic systems - affecting Perceived Usefulness. However, it can logically be assumed that an independent variable that negatively affects a
19 person’s ability to use a given system will act as a moderator in predicting that person’s behavioral intention to use. Thus, it is hypothesized that:
H3: There is a positive relationship between PU and BIU that is moderated by Health
Issues, such that an increase in Health Issues will weaken this relationship.
Van Der Heijden’s (2004) research into technology adoption models posited that joy serves as the main intrinsic motivator of people leading to their behavioral intentions to use of a certain technology product. “Joy, or perceived enjoyment, is the extent to which using a system is perceived to bring pleasure and fulfillment for <one’s> own sake, apart from any anticipated performance consequences” (Lowry et al, 2012; Hong & Tam, 2006; Lee et al., 2005). Thus, frustrations that might arise from a difficulty to use and interact with a system will lead to a decreased sense of enjoyment; and vice versa. Thus, it is hypothesized that:
H4: There is a positive relationship between PEOU and Joy.
Motivational theory indicates that individuals engage with systems for one of two reasons: to obtain an outside benefit that is external to the user-system interaction - such as improving job performance - or to gain a benefit that is internal to the user-system interaction – such as enjoying a pastime or to feel rejuvenated. “The definition of perceived enjoyment specifies the extent to which fun can be derived from using a system” (Van der Heijden, 2004). In such cases of intrinsic motivation, it can be logically argued that a system which satisfies the desire to experience joy will lead to a user intending to use it again in the future. Thus, it is hypothesized that:
H5: There is a positive relationship between Joy and BIU.
Many VR industry-experts have quoted health issues (nausea, dizziness, headache) as being the primary barrier to the adoption of VR (Newell & Valve Employees, 2017). Though there is a research gap within publicly available studies addressing the area of VR health issues affecting individuals’ Perceived Joy, it is logical to assume that an independent variable that affects person’s physical well-being and ability to
20 use a given system will act as a moderator in predicting that person’s behavioral intention to use. Thus, it is hypothesized that:
H6: There is a positive relationship between Joy and BIU that is moderated by Health
Issues, such that an increase in Health Issues will weaken this relationship.
In HMS, Perceived Ease of Use is a baseline expectation (Sweetser & Wyeth, 2005). If PEOU is not present (i.e. a game is difficult to control and use), players will experience frustrations and grow dispirited or even uninterested to continue with the game. This is not to say that a game should be easy to complete, but rather that it should not be cumbersome and frustrating to play. Ease of use allows a player’s attention to be focused on exploring the game and finding exciting experiences to participate in, as well as discovering the “available possibilities” for interaction with the game (Sun & Zhang, 2006). Previous studies by researchers such as Rouibah (2008) among others, found strong correlations between PEOU and Curiosity. Combining these findings with the observations that lowering frustrating complexity removes potential barriers to use, it can be argued that a system/game with high Ease of Use better enables exploratory behavior, thus giving players more opportunities to engage in novel interactions. These experiences will likely lead to an increase in a player’s curiosity about a system (Posner & Boies, 1971). As such, it is hypothesized that:
H7: There is a positive relationship between PEOU and Curiosity.
Curiosity is a main motivational driver that increases engagement with a subject, as well as drives further exploratory behavior (Kashdan, Rose, & Fincham, 2004). This exploratory nature invites further discovery, a process that brings excitement to the user (Reiss, 2004). Consequently, curiosity can exist without the requirement that Joy be present as well. However, curiosity does amplify excitement about an interaction with a system (Sweetser & Wyeth, 2005) – leading to a user wanting to repeat that engagement to re-live the excitement over again (Kashdan, Rose, & Fincham, 2004). This forms the basis of the following hypothesis:
21
4.0 Method
A survey analysis will be carried out to judge consumers’ willingness to adopt VR as a new platform into their lives. This will be performed in the form of a live demonstration of a VR system in action. A questionnaire probing for impressions of this demo was chosen as the research method for this study. Survey questions will be administered before and after the live VR demo. These include questions that test the relevance of the various variables of the HMSAM model, as well as additional criteria that serve to judge whether any new characteristics have arisen due to the changing technology landscape.
4.1 Sample
A sample of 90 university students (70 male, 20 female) aged between 21-27 (mean 24.05, SD 1.87) voluntarily participated in the experiment. In an attempt to represent a wider age group, this was further supplemented by an additional 15 full-time employed participants (10 male, 5 female) aged between 30-51, thus bringing the total number of participants to N=105.
22 The selection procedure for the participants was limited by the constraints imposed on the demo by the Virtual Reality setup: It required a controlled, unobstructed space of at least 3m x 3m with specialty VR equipment set up and calibrated in advance. Due to the relatively lengthy time required for a single demo (minimum 15–30 minutes per participant), the space also had to be secure and lockable to allow for multiple days of demonstrations without change to the setup. As such, an empty college dormitory room was selected as the location of the demo. This provided access to a diverse sample of subjects originating from a varied set of nationalities. However, it also limited a large portion of the demographics to bachelor and master’s students of the University of Amsterdam and Vrije University. The older 30+ group was characterized by visiting family and friends of residents of the student dormitory, and were all employed full-time. As such, the sample is limited in its representation of the general population.
Subjects were recruited to participate in the experiment through various means:
1. Flyers around the building detailing the opportunity to try out VR 2. Posts on the electronic online messaging boards of the dormitory 3. Knocking on doors and asking people in person to participate 4. Open door policy for anyone walking by
5. Word of mouth and referrals
Due to the voluntary participation, there is a natural tendency for those to participate who are at least somewhat interested in VR. However, this is controlled for through a survey question that asks, “how curious are you about VR?” A score of 4 in this case indicates neutral and acts as the baseline for someone who is neither curious about - nor hates the technology. With the marketing push for VR over the last year, it is to be expected that at least some people have developed a curiosity toward the technology, and the sample’s average 4.4 signifies that this is indeed the case and that most people are between neutral and somewhat agree in their level of curiosity for VR.
23
4.2 Measure
4.2.1 Materials
A survey was constructed consisting of 41 questions in total (see appendix) that were split into two groups to be answered pre, and post-demo. The pre-demo section addresses basic demographic information such as Age, Gender and Occupation (among others), as well as questions pertaining to participants’ affinity towards technology and their previous exposure to VR; along with their curiosity about VR. The post-demo section on the other hand contains questions aimed at testing the variables of the model discussed above. A seven-point Likert scale was used, ranging from 1=strongly disagree to 7=strongly agree. Some additional scales were included in the questionnaire for the sake of completeness, however these will not be featured in the explicit discussion of the model at hand.
The scales used were taken from previous studies dealing with past TAM and HMS models from which they were adapted to a VR context. They have been internally proven to be reliable, and content and construct validity has also been established for them. The scales address the following variables:
Table 1: Scales and example items
Construct Item examples Notes
Perceived Ease of Use
PEOU1. My interaction with VR was clear and understandable.
PEOU2. Interacting with VR did not require a lot of my mental effort.
Based on the original scales used in the HMSAM model from (Lowry et al) and modified to a VR context
Perceived Usefulness PU2. VR helped me better pass time. PU5. VR helped me feel rejuvenated.
Perceived Joy
JOY1. I found playing in VR to be enjoyable.
JOY5. The VR experience was pleasurable.
Curiosity
CUR1. This experience excited my curiosity.
CUR3. This experience aroused my imagination.
HLTH2. I felt sick from my interaction with VR (nausea/dizziness/etc.)
Adapted to a VR context from the Stanford Medicine Health
24 Health Issues HLTH3. I did not experience negative
physical effects from using VR
Research Instruments list of scales (Stanford Medicine, 2017).
One additional scale – the ‘Health’ variable – has been constructed and added following many citations of its effects on VR. Internal reliability analysis was conducted. See appendix for a full breakdown of scales and items.
4.2.2 Apparatus
The full HTC VIVE kit (version 1) was used as the virtual reality demonstration unit. It made use of ‘room-scale’ in a 3x3m area with the base stations situated at a 2.5m height on opposing sides of the space to track the players’ movements. Both of the two hand-tracking controllers were employed.
A high-end custom-built PC was used to drive the VR demos. The internal hardware consisted of an Intel Core i7 5770K processor paired with a Nvidia GTX 1080 GPU and 32GB of DDR4 RAM, with the demos locked to a minimum 90 frames per second, running off a 500GB SATA3 Samsung SSD. The headphones of the Vive (and consequently the 3D directional sound) were disabled during the demos – instead the sound was played through a 5.1 surround sound system in order to be able to communicate with and instruct the participants. This allowed other participants waiting for their turn to hear the action, which was further aided by a projector displaying the VR player’s internal point of view on an extra external screen for everyone to watch and follow on. (see image below).
Figure 10: Live VR demo
25 Three demos were shown to each participant. These consisted of three different, unique experiences that demonstrated various aspects of VR. They were chosen in a manner that attempted to control for genre specific biases.
The first demo was the Archery exhibit from ‘The Lab.’ This provides an ideal introduction to VR, allowing players to familiarize themselves with the movement and controls in VR, while also providing an interesting, dexterous interaction with the world.
The next exhibit is Google Earth’s VR application. The welcome tutorial was run for participants. This delivers a very different experience to the previous one. Interaction is much more limited; however it allows players to explore the entire planet and fly around countries and landscapes and cities in 3D. This caters to a different type of curiosity in people. Those who do not necessarily enjoy video games and digital interaction are also able to find enjoyment in this demo, and often participants lose track of time trying to find their house or other familiar places.
The ‘Waltz of the Wizard’ demo was chosen as the final experience due to it being the most complex one. By this time, participants should have gotten accustomed to interacting with VR. This demo brings everything together: moving around within the 3x3m space of the room and interacting with virtual objects within the VR demo.
4.3 Procedure
Subjects were given a questionnaire and asked to fill in the pre-demo section of the paper. Following this, the VR headset was placed on their faces and the straps adjusted for comfort and head size. Participants were asked to view a text within the VR headset and adjust the sharpness through a calibration knob until they got the clearest image possible. Following this, an empty virtual orientation room was loaded up and participants were asked to walk around and explore the edges of the real 3x3m space. The hand-tracking controllers were then given to participants and they were asked to touch the edges/walls of
26 the space to familiarize themselves with the digital ‘chaperone’ system that indicates the correlating wall boundaries within the VR world.
Next, the Archery demonstration was loaded. Subjects were told to look around and pick up the bow next to them. No further instructions were given, instead subjects were left to their own devices to figure out the controls using intuition. This is a crucial part to arousing immersion and curiosity. After the familiarization and some target practice, one round of castle defense was played before moving on to the next demo.
In Google Earth VR, participants were shown the welcome tutorial of the app in which they solely an onlooker. Once this is over, instructions were given to participants in how to interact with the 3D Earth, and as players were once again left to themselves to explore as they pleased.
In the final demo of ‘Waltz of the Wizard,’ subjects were completely left on their own to explore and figure out the environment. The demo features an assortment of activities to perform in a wizard’s castle, allowing the player to cause complete mayhem through interaction with a wide range of magical objects and the ability to mix potions of a number of ingredients. If players were struggling to figure out things to see/do, some pointers and advice were given on where to look or what to touch.
In total, 15-30 minutes were given to each participant to experience the demos. This depended on how fast subjects were able to pick up and understand the mechanics of the interactions. The goal was to facilitate players to intuitively explore and figure out the controls and interaction VR on their own, not to strictly control for time. Following the completion of the demos, participants were asked to fill in the remainder of post-demo section of the questionnaire.
27
5.0 Results
5.1 Descriptive Statistics
Table 2: Descriptive StatisticsDescriptive Statistics
N Minimum Maximum Mean Std. Deviation
Age 105 21.00 51.00 27.333 8.414
Gender (1=Male, 0=Female) 105 0 1 .71 .454
Nationality (1=Dutch, 0=Other) 105 0 1 .286 .453
Affinity to tech 105 2.00 7.00 5.476 1.474
Capability to spend 105 3.00 7.00 5.476 1.056
Plan Adoption of VR (pre) 105 2.00 7.00 3.809 1.569
Perception Change 105 4.00 7.00 6.000 .820
Plan Adoption of VR (post) 105 2.00 7.00 4.714 1.206
PEOU 105 3.43 6.29 5.176 .703 PU 105 4.00 6.50 5.119 .608 BIU 105 2.00 7.00 4.476 1.194 JOY 105 4.67 7.00 6.182 .553 CURI 105 4.67 7.00 6.158 .666 HLTH 105 1.00 6.33 2.206 1.718 Valid N (listwise) 105
28
5.2 Correlation Analysis
Table 3: Correlations
Owing to the nature of the VR topic and the participants’ general enthusiasm for the technology, many of the variables exhibit correlations with each other (both at the 1% and 5% significance levels). However, among those surveyed there are some notable (and unexpected) correlations: age and ‘affinity towards technology’ are moderately correlated (0.365**); for every year increase in age, there is a 0.365-point increase in affinity to tech. On the other hand, ‘Age’ is not correlated with ‘Capability to Spend on Technology’ as could be expected. ‘Gender’ is highly correlated with ‘Perceived Joy’ at 0.655, as is ‘Behavioral Intention to Use’ at 0.485. Of interest regarding BIU is its negative correlation (-0.401) with ‘Nationality,’ indicating that the Dutch portion of the sample are less likely to ‘Intend to Use’ VR compared to the ‘non-Dutch’ portion. No obvious differences between the two subgroups were found. Finally, the factor ‘Health Issues’ demonstrated significant, mostly negative correlations across the board. This is in
Correlations 1 2 3 4 5 6 7 8 9 10 11 1. Age - 2. Gender .352** - 3. Nationality .302** .167 - 4. Affinity to tech .365** .349** .010 - 5. Capability to spend .171 .315** .014 .131 - 6. Perception change .383** -.129 -.129 -.358** .277** - 7. PEOU -.087 .246* -.117 .488** .006 -.452** (.76) 8. BIU .335** .253** -.401** .125 .098 .343** .297** (.74) 9. JOY -.049 .655** -.114 .030 -.273** .141 .269** .485** (.78) 10. CURI .376** .151 -.204* -.306** .142 .410** .106 .172 .391** (.75) 11. HLTH .376** .725** -.261** -.229* .219* -.057 -.265** -.384** -.668** -.173 (.96)
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).
29 line with expectations. The strong positive correlation of ‘Health Issues’ with the ‘Gender’ dummy variable (correlation of 0.725 when male=1) are likely explainable due to the low number of females represented in the sample.
5.3 Collinearity VIF Analysis
Table 4: Collinearity and VIF analysis
Coefficients Independent Variables Collinearity Statistics Tolerance VIF 12. Age .543 1.842 13. Gender .288 3.468 14. Nationality .670 1.492 15. Affinity to tech .414 2.418 16. Capability to spend .615 1.625 17. Perception change .390 2.561 18. PEOU .407 2.458 19. JOY .320 3.121 20. CURI .487 2.051 21. HLTH .279 3.584
Dependent Variable: BIU_TOT
Following the observation of high correlations demonstrated above, a variance inflation factor test (VIF) was performed to quantify the severity of possible multicollinearity between the independent variables when taking BIU as the dependent variable. Among all the explanatory factors, only Gender, JOY and HLTH displayed mild signs of collinearity (slightly above the VIF threshold of 3). HLTH and JOY form an integral part of the study’s model, but Gender has been excluded from being a potential control variable in further analysis.
30
5.4 Reliability tests
Table 5: Reliability statistics
Reliability Statistics
Scale Cronbach's Alpha N of Items
Perceived Ease of Use (PEOU)
0.756
6*
*Perceived Usefulness (PU)
0.193
4
Behavioral Intention to Use (BIU)
0.736
3
Perceived Joy (JOY)
0.779
6
Curiosity (CURI)
0.752
3
Health Issues (HLTH)
0.985
3
The Perceived Ease of Use, Behavioral Intention to Use, Perceived Joy and Curiosity scales all have a relatively high level of reliability, with Cronbach’s Alpha scores around 0.75. However, one item (PEOU3) of the PEOU scale only had a correlation of 0.134 with the total score of the scale. This item probed participants’ impressions of the VR experience being “trouble free,” an arguably ambiguous phrasing; thus, dropping this item raised the scale’s Cronbach Alpha score from 0.734 to 0.759. Perceived Usefulness on the other faced more severe problems. Even when dropped, none of the items would substantially increase the reliability of the scale to acceptable levels. This is in contrast to other adoption models (e.g. Technology Acceptance Model), where Perceived Usefulness is traditionally one of the most reliable and stable independent factors. However, in this instance it is not, and as such it will be excluded from further analysis in this study. Health Issues on the other hand exhibits a very high reliability with Cronbach Alpha = 0.985, with total correlation >0.95 for all its three items.
31
5.5 Regression Analysis
H4: PEOU JOY
Table 6: H4 Regression analysis
R R² R² Change B SE β t Sig. Model 1 .049a .002 AGE -.003 .006 -.049 -.501 .618 Model 2 .269b .072 PEOU .212 .075 .269** 2.834 .006 Model 3 .270c .073 .071 AGE -.002 .006 -.026 -.273 .785 PEOU .210 .075 .267** 2.787 .006 Dependent Variable: Perceived Joy Note: Statistical significance *p <.05; **p <.01; ***p <.001
a. Predictors: (Constant), AGE b. Predictors: (Constant), PEOU c. Predictors: (Constant), AGE, PEOU
A hierarchical regression analysis was carried out to examine the ability of Perceived Ease of Use to predict levels of Perceived Joy, after additionally controlling for age. As the first step of the hierarchical regression, age was included as a predictor. This model was not statistically significant. Including Perceived Ease of Use (model 3) however was significant; β = 0.264 with p < .05 and explained 7.3% of the total variance in perceived joy. In other words, if a person’s Perceived Ease of Use increases by one, their Perceived Joy will increase for 0.267.
32
H5: JOY BIU
Table 7: H5 Regression analysis
R R² R² Change B SE β t Sig. Model 1 .335a .112 AGE .048 .013 .335*** 3.606 .000 Model 2 .485 b .235 JOY 1.046 .186 .485*** 5.627 .000 Model 3 .603c .364 .252 AGE .051 .011 .360*** 4.548 .000 JOY 1.084 .170 .503*** 6.358 .000
Dependent Variable: BIU Note: Statistical significance *p <.05; **p <.01; ***p <.001
a. Predictors: (Constant), AGE b. Predictors: (Constant), JOY c. Predictors: (Constant), AGE, JOY
A hierarchical regression analysis was carried out to inspect the ability of Perceived Joy to predict levels of Behavioral Intention to Use, after additionally controlling for age. As the first step of the hierarchical regression, age was examined as a predictor. This model was statistically significant with a positive β 0.335 at p < .001, and explained 11.2% of the variance in Behavior Intention to Use. After entry of Perceived Joy at Model 3, the total variance explained by the model as a whole was 36.4%; p < .001. The introduction of Perceived Joy explained an additional 25.2% of the variance Behavioral Intention to Use, after controlling Age. In the final model, both predictor variables were statistically significant, with JOY recording a higher Beta value (β = 0.503, p < .001) than Age (β = 0.360, p < .001). In other words - given a stable age - if a person’s Perceived Joy increases by one, their Behavioral Intention to Use will increase by 0.364.
33
H6: PEOU CURI
Table 8: H6 Regression analysis
R R² R² Change B SE β t Sig. Model 1 .376a .141 AGE .030 .007 .376*** 4.119 .000 Model 2 .106b .011 PEOU .100 .093 .106 1.077 .284 Model 3 . 401c .161 .019 AGE .031 .007 .388*** 4.263 .000 PEOU .132 .086 .139 1.529 .129
Dependent Variable: CURI Note: Statistical significance *p <.05; **p <.01; ***p <.001
a. Predictors: (Constant), AGE b. Predictors: (Constant), PEOU c. Predictors: (Constant), AGE, PEOU
A hierarchical regression analysis was carried out to inspect the ability of Perceived Ease of Use to predict levels of Curiosity, after additionally controlling for age. As the first step of the hierarchical regression, Age was examined as a predictor. This model was statistically significant with a positive β = 0.376 at p < .001, and explained 14.1% of the variance in Curiosity. After entry of PEOU at Model 3, the total variance explained by the model as a whole increased to 16.1%; p < .001. The introduction of PEOU explained an additional 1.9% of the variance in Curiosity, after controlling Age. In the final model, only Age as a predictor variable was statistically significant (β = 0.388; p < 0.001), while PEOU was not. In other words - given a stable age - if a person’s Perceived Ease of Use increases by one, it has no statistically significant influence on Curiosity.
34
H7: CURI BIU
Table 9: H7 Regression analysis
R R² R² Change B SE β t Sig. Model 1 .335a .112 AGE .048 .013 .335*** 3.606 .000 Model 2 .172b .030 PEOU .309 .174 .172 1.777 .079 Model 3 .339c .115 .097 AGE .045 .014 .314*** 3.127 .002 PEOU .097 .180 .054 .539 .591
Dependent Variable: BIU Note: Statistical significance *p <.05; **p <.01; ***p <.001
a. Predictors: (Constant), AGE b. Predictors: (Constant), CURI c. Predictors: (Constant), AGE, CURI
A hierarchical regression analysis was carried out to inspect the ability of Curiosity to predict levels of Behavioral Intention to Use, after additionally controlling for age. As the first step of the hierarchical regression, Age was examined as a predictor. This model was statistically significant with a positive β = 0.335 at p < .001, and explained 11.2% of the variance in BIU. After entry of PEOU at Model 3, the total variance explained by the model as a whole only increased by 0.97% to 11.51% (p < 0.01). The introduction of CURI had no significant effect on BIU, after controlling Age. In the final model, only Age as a predictor variable was statistically significant (β = 0.314; p < 0.01), while CURI was not. In other words - given a stable age - if a person’s Curiousness increases by one, it has no statistically significant influence on Behavioral Intention to Use.
35
5.6 Moderation
H8: JOY BIU (Moderated by HLTH)
Table 10: H8 Moderation analysisCoefficient SE t p Intercept i1 2.942 .281 10.469 .0000 JOY (X) c1 .608 .172 3.539 .0006 HLTH (M) c2 -.486 .057 -8.507 .0000 JOY*HLTH (XM) c3 -.159 .032 -4.930 .0000 AGE .089 .006 14.884 .0000 R²=0.4947 p<0.0000 F=144.4387
Conditional effect of X on Y at values of the moderator(s): Table 11: H8 Moderation analysis cont.
Effect SE t p 1.000 .449 .145 3.092 .0026 1.000 .449 .145 3.092 .00026 1.667 .343 .129 2.660 .0091 2.000 .290 .121 2.389 .018 5.333 -.240 .093 -2.588 .011
The regression coefficient for XM is c3= -0.159 and is statistically different from zero at p<0.001. As such, it can be concluded that the effect of Joy on the behavioral intention to use the VR system is related to the health issues that one might experience while using VR. The negative effect of Health Issues on behavioral intention to use (alone c2= -0.486 at p=0.0000) are able to overcome the positive effect that Joy alone has on BIU (c1=0.608 with p=0.0006). Furthermore, this model accounts for 49.47% of the variance in Behavioral intention to use. A further inspection of the conditional effects indicates that the relationship between Joy and behavioral intention to use turns turns negative once Health Issues reach a level between 2.000 and 5.333 (effect= 0.56, SE = 0.09, CI: .371 to .745. Moreover, it can be observed from further
36 investigating the interactions that the slope of Health Issues is constantly negative, indicating that this factor this is a trend regardless of the level that ‘Health Issues’ exhibits.
Moderated Mediation
Table 12: Moderated mediation analysis
MODEL SUMMARY
R R2 P
0.7255 0.5264 0.0000
ANTECEDENT
CONSEQUENT
Perceived Joy (M) Behavioral Intention to Use (Y)
Coeff. SE P Coeff. SE P PEOU .210 .075 .000 .316 .123 .011 JOY - - - 1.034 .279 .000 HLTH - - - -1.324 .480 .007 JOY-HLTH - - - -.290 .086 .001 AGE -.002 .006 .785 .090 .012 .000
The above tables express the model’s predictive capabilities as a whole. Those relationships that displayed significance in the previous analyses have been included: that is PEOU JOY BIU, moderated by HLTH. Alone, Perceived Ease of Use explains 8.80% of the variance in of Behavioral Intention to Use, while taken together with JOY and HLTH, this number increases to 52.64% as indicated by R2. JOY has a β coefficient of 1.034, while HLTH is β = -1.324. They are both significant (p < 0.001 , p < 0.01).
6.0 Discussion
The aim of this study was to Investigate which factors of existing technology acceptance models (specifically the HMSAM - Hedonic Motivation System Adoption Model) are still applicable today, and whether they need to be amended in order to better explain modern innovation adoption trends; with regards to new technologies such as Virtual Reality.
37 From the analysis of the HMSAM model and further literature, it was expected that there would be a positive relationship between Perceived Ease of Use and Behavioral Intention to Use, mediated by three key variables: Perceived Usefulness, Perceived Joy and Curiosity. These expected relationships motivated the majority of the individual hypotheses, with the addition of Health Issues being included as a moderator between the mediator variables (PU, JOY, CURI) and the final BIU dependent variable.
In terms of the questionnaire items, they were split into two groups (1) addressing demographic control variables in a pre-demo section (2) addressing hypothesis testing with targeted scales in a post-demo section. The descriptive statistics of the results are shown in Table 2 in the section above, and of special interest is the relatively high mean value (6 on the seven-point Likert scale) for Perception Change (as a result of participating in and experiencing the VR demo). This reinforces Newell’s (2017) statement that “VR cannot be explained merely through words. It has to be personally experienced.” This is underpinned by the increase in Planned Adoption of VR between the pre, and post-demo section of the questionnaire. Furthermore, from the qualitative feedback collected throughout the demos, participants often exclaimed that they ‘logically knew what they were about to experience, however actually living through immersive demo was completely different - much more powerful than they had anticipated.’ Just like Carmack (2014) had also explained, VR seems to impact people more on an emotional level rather than just a logical one, and hence the overwhelming positive reception of the technology.
In terms of the actual hypothesis testing, the results show that for the specific sample represented in this study, only the mediator Perceived Joy had a statistically significant effect in explaining the variance in Behavioral Intention to Use. Specifically, this is the mediated pathway between Perceived Ease of Use and Behavioral Intention to Use. Hence, hypothesis H4 and H5 are supported. In other words, there is indeed a positive relationship between PEOU and JOY, as well as between JOY and BIU.
Curiosity as a mediator between PEOU and BIU on the other hand was not support by the results. Neither hypothesis H7 (There is a positive relationship between PEOU and Curiosity) nor H8 (There is a positive relationship between Curiosity and BIU) had a large impact or was statistically significant, with
38 – or without – age as a controlling variable. In other words, there was no statistical support of the existence of a positive relationship between PEOU and CURI, or CURI and BIU as suggested by the original HMSAM model by Lowry et al (2012). While research is lacking in the field regarding this concrete result, a possible explanation comes by way of the qualitative observations made of the participants: In many cases, participants coming out of VR had a look of wonder and amazement on their faces and exclaimed in enthusiasm and curiosity for the system. Their remarks had a noticeable pattern to them: “I’m really curious about the future and how VR will integrate into various aspects of it.” These remarks were mostly positive; however, they did highlight a key effect of curiosity; it was more about a long-term interest about the technology and its uses, rather than a short-term immediate intention to adopt and use.
Hypothesis analysis regarding Perceived Usefulness was not possible to carry out, as reliability testing indicated a major concern with the scale. In contrast to JOY and CURI, the scale of PU returned a sub 0.2 alpha score, which was not significantly improved by dropping any of the items. A factorial analysis revealed at least two, but possibly three underlying factors in the 4-item scale. As such, further testing of PU was halted, and evidence in favor of, or against hypothesis H1 (There is a positive relationship between PEOU and Perceived Usefulness) or H2 (There is a positive relationship between Perceived Usefulness and BIU) could not be established.
Such a low alpha score for PU is contrary to expectations, as in other technology adoption models (e.g. Technology Acceptance Model), PU is usually one of most reliable and stable independent factors. However, there are reasons why this outcome might not be a coincidence. Traditional TAM models usually take an extrinsic motivational approach when dealing with PU. In such a case, PU does indeed prove to be a very stable and reliable independent factor. However, the original HMS (and by extension the HMSAM) model tries to address adoption factors from a hedonic motivational perspective, and thereby aims to use intrinsic motivations to explain its variables, including PU.
39 The PU items that have been adopted from the HMSAM model and adapted to this study are:
Table 13: PU Scale items
1. PU1. Being in VR decreased my stress. 2. PU2. VR helped me better pass time. 3. PU3. VR provided a useful escape. 5. PU4. VR helped me feel rejuvenated.
These items were useful and reliable for the original HMSAM model. However, that study examined users' reactions to a very simplistic, stress-free video game; whereas using them in a VR context arguably changes this situation (also based on the verbal feedback of the participants). This is because the experience is an immersive one that causes a much greater emotional impact. For example, participants who had vertigo claimed to experience a similar, real stress and fear within VR, something that was initially not anticipated. This caused them to answer certain questions very differently to others who didn't have vertigo (or a number of other reactions). There were other examples of such opposing responses to the PU scale, but in essence, 'usefulness' in a VR context is a more subjective concept than in previous studies. As such, this scale and the items within it proved to be less reliable in measuring ‘usefulness’ as the underlying factor.
To test hypothesis H3 (there is a positive relationship between PU and BIU that is moderated by Health Issues, such that an increase in Health Issues will weaken this relationship) and H6 (there is a positive relationship between Joy and BIU that is moderated by Health Issues, such that an increase in Health Issues will weaken this relationship), Health Issues was individually examined as a moderator between the dependent variable Behavioral Intention to Use and the three independent explanatory variables PU, JOY and Curiosity. A negative moderation effect was expected, as experiencing health issues such as nausea, dizziness and headaches due to interactions with VR would hinder all other factors as described in the theory development section above. Owing to the low reliability of the PU scale, it was excluded from this analysis. Nevertheless, the relationships of Perceived Joy and Curiosity were tested,
40 and Health Issues did prove to exhibit a statistically significant negative moderation effect on relationship between the JOY and Behavioral Intention to Use. In fact, this negative effect was nearly substantial enough to overpower the positive relationship linking JOY and BIU. As for the moderation effect of Health Issues on the relationship between CURI and BIU, no statistically significant results were observed. In other words, while hypothesis H6 is supported (H6: Health Issues has a negative moderation effect on the relationship between JOY and BIU).
A wholesome overall test of the model with all variables included in a ‘moderated mediation’ analysis reveals the recurrent theme of this study: the pathway between PEOU and BIU that is mediated by JOY proves to be a positive and statistically significant relationship, and is further improved by the moderation of HLTH. That is: an increase in PEOU leads to an increase in JOY, which in turn again increases the rate of BIU of an individual, given that they do not experience a high level of Health Issues. The same is not valid for the pathway including Curiosity.
7.0 Limitations & Future Research
The obvious main limitation of this study is the absence of a reliable scale to measure participants’ Perceived Usefulness. Without such a scale, it is impossible to say – based on this sample – whether or not the portion of the HMSAM model involving PU is supported. Future research into this topic should research the motivations that determine subjects’ perceived usefulness within a VR context, combining them into a reliable scale with which PU can be measured; and ultimately the full model evaluated.
Due to the complexities associated with the setting up and running of a Virtual Reality demo, the method of convenient sampling was employed to gather the subjects to participate in the study. This undoubtedly introduced a number of sampling biases including: curiosity bias, gender bias, age bias, and occupational bias (among others). Though age was controlled for within the analyses, gender and
41 occupation could not due to their low sample sizes. Future research should also explore the impact that nationality has on the results; to determine whether the significant negative correlation between Dutch and Behavioral Intention to Use was due to a mere sampling error or an actual underlying relationship. The above limitations would mostly be addressed by a larger sample size chosen picked through true random sampling; ensuring a more representative overview of the actual population.
Finally, based on the qualitative feedback of the participants, the high level of immersion achieved in VR greatly aided in the enjoyment of the experience. This is in line with comments made by Carmack (2014) of Oculus VR; in which he states that “virtual reality has to be experienced to be understood.” In contrast to the HMSAM model which identified Immersion as an outcome of Joy, Curiosity and Control, research focusing on immersion as an independent explanatory variable affecting Perceived Joy would serve as a valuable extension of the study.
8.0 Conclusion
With regard to the research question: which factors of the HMSAM model influence consumers’ willingness and intention to adopt Virtual Reality devices into their daily lives, the findings of this study partially address the research question: While it cannot be determined whether Perceived Usefulness is a valid determinant of Behavioral Intention to Use, according to the sample, Perceived Joy exhibits a significant and positive effect on the relationship, whereas Curiosity does not. Furthermore, Health Issues displays a significant negative effect as a moderator, and should be included in the model. With respect to the HMSAM model, it seems that the alterations it made to the original HMS model by Van der Heijden (2004), are not supported. Therefore, the final suggestion of this study is that the original HMS model is the more relevant model to VR (and possibly new technologies in the future), but should be amended with Health Issues as a moderator.
