University of Amsterdam
Master Thesis
Amber van der Neut (10565027)
MSc. Business Administration - Marketing Track
Advertising in the Age of Artificial Reality
An Experimental Research: The Effects of 360° Advertisements on Consumer BehaviourStatement of Originality
This document is written by student Amber van der Neut who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are 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.
Table of Contents
1. Abstract ... 4
2. Introduction ... 5
3. Literature Review ... 8
3.1 Antecedents of Virtual Reality Effects ... 9
3.2 Virtual Reality in other Industries ... 12
3.3 Virtual Reality in Advertising ... 14
3.4 Purchase Intention ... 17 3.5 Perceived Risk ... 19 3.6 Product Experience ... 21 3.7 Brand Evaluation ... 22 4. Methodology ... 23 4.1 Research Strategy ... 24
4.2 Sample and Procedure ... 24
4.3 Manipulation ... 25 4.3.1 Manipulation Check ... 25 4.3.2 Stimuli ... 25 4.4 Survey ... 28 4.4.1 Perceived Risk ... 28 4.4.2 Purchase Intention ... 28 4.4.3 Brand Evaluation ... 29 4.4.4 Control Variables ... 29 5. Results ... 30 5.1 Manipulation Check ... 30 5.2 Hypothesis Testing ... 31 6. Discussion ... 42 6.1 General Discussion ... 42 6.2 Theoretical Implications ... 43 6.3 Managerial Implications ... 46
6.4 Limitations and Future Research ... 48
7. Conclusion ... 50
8. References ... 52
Appendix A. Manipulation Check ... 59
Appendix B. The Survey ... 60
List of Tables and Figures
Tables
Table 1. Online experiment: 2X2 full factorial between-subjects design ... 27 Table 2. Correlations, means and standard deviations ... 32 Table 3. Means and standard deviations on purchase intention ... 34 Table 4. Results regression analysis of effect advertisement type on perceived risk and purchase intention ... 37 Table 5. Results regression analysis of direct effect advertisement type on purchase
intention and indirect effect on purchase intention through the mediator perceived risk ... 37 Table 6. Results regression analysis of effects product type, advertisement type and product type * advertisement type on perceived risk ... 39 Table 7. Means and standard deviations of the conditions split into two product type
dimensions on perceived risk ... 39 Table 8. Summary of results ... 41
Figures
Figure 1. Antecedents of virtual reality effects ... 11 Figure 2. Conceptual model ... 23 Figure 3. Unstandardized regression coefficients model ... 38 Figure 4. The moderating role of product type on the relationship between the
1. Abstract
Virtual reality (VR) provides a gateway for marketers to innovatively reach consumers with their marketing communications. Since marketers are essentially responsible for driving the customer experience, they stand to make huge achievements with VR. VR advertisements enable consumers to experience products realistically in a virtual world, thereby mitigating the problems associated with the lack of physical contact with products. Although the employment of VR in advertising has increased by certain innovative brands, there is little understanding about its impact on consumer behaviour. It is therefore important that this study addresses the gap by examining to what extent the most frequent use of VR in advertising, which is the 360° product experience, has a positive effect on the purchase intention amongst consumers. Additionally, this study investigates whether the perceived risk of making a purchase mistake explains this relationship. Primary data was obtained for this study through an experiment embedded in a survey. The research looked at how 197
participants responded differently to VR advertisements as compared to 2D advertisements of the same product. The results showed that participants had a higher intention to purchase the advertised product after being exposed to the VR advertisement than the participants exposed the 2D advertisement. A lower level of perceived risk to make a purchase mistake explained this relationship. As such, the strategic potential of VR in marketing communications is demonstrated by this study. It is highlighted that this new age of advertising enables marketers to provide cutting edge experiences for their increasingly sophisticated consumers in order to encourage them to bond with their brand.
2. Introduction
On the 25th of February 2016 Facebook CEO Mark Zuckerberg hosted a live chat where he and Facebook’s director of Artificial Intelligence Yann LeCun substantiated their investments in Virtual Reality (VR) the upcoming years. They expect that the next 10 years will be ‘The Decade of VR’, in the same way the smartphone has defined technology over the last 10 years (Smith, 2016). YouTube CEO Susan Wojicicki has shown her excitement about virtual reality as well and made it YouTube’s new frontier to empower creators to build immersive, 360° content (Tabaka, 2015).
While technological innovations continue to impact digital marketing, VR has taken a prominent place in the marketing landscape over the last few years by redefining the way we are experiencing things. VR refers to an artificial, computer-generated environment that uses high-end graphics and audio sensations to make users feel as if they are in a real world (Clark, 2017). It creates a strong feeling of being physically present in a virtual world. This research will focus on 360° advertisements through which you can explore the virtual environment in all directions. Although the use of the term ‘virtual reality’ to describe 360° advertisements has been disputed because of the lack of interaction, it is argued that the immersive nature of the virtual environment and the realism achieved do belong to the VR category (Swann, 2001). This type of VR would fall under Desktop VR, where the user views the VR on a standard computer and is able to interact with the virtual world in the advertisement by using a standard mouse (Swann, 2001).
The immersive environment and realism these videos can offer, show some great opportunities according to Zuckerberg: “When we can’t experience things in real life, when we run up against the limits of reality, VR is going to make our reality that much better” (Smith, 2016). Accordingly, VR experiences introduce huge implications for the digital marketing and advertising landscape. It could be that marketers, who are essentially
responsible for driving the customer experience, stand to make huge achievements with VR advertising now that selling experiences has become a crucial part of success (Crabtree, 2016). As a matter of fact, according to Marketo (2017) 30% of Forbes Global 2000 (the world’s biggest public companies) consumer-focused companies will experiment with VR or augmented reality in 2018.
Despite these innovative companies, virtual reality remains a relatively unexplored area in the consumer behaviour literature and therefore the use by marketers has been limited so far (Suh & Lee, 2005; Tussyadiah, Wang, Jung, & Dieck, 2018). Empirical evidence about VR from various other fields of study, including education, healthcare, entertainment and tourism demonstrate that VR exposure leads to positive behavioural outcomes, such as
product recall, brand recognition, and memory of experiences (Kim & Biocca, 1997; Mania & Chalmers, 2001). The question for marketers remains if these positive effects of VR also apply to the field of advertising.
For the field of advertising, despite the employment of VR in digital advertisements is slowly but surely increasing, its impact on consumers has not been explored extensively (Tussyadiah et al., 2018). Advertising is shifting into an age of artificial reality, but research about the effectiveness of these VR advertisements is lacking. Existing studies that are relevant for this shift in advertising mainly focus on (1) 3D advertising and (2) informational VR messages in other industries (Kerrebroeck, Brengman, & Willems, 2017). First, previous work about 3D advertising has shown some indication that consumers respond favourable to virtual product experiences, considering that these experiences enable them to obtain more information about the product (Debbabi, Daasi, & Baile, 2010; Suh & Lee, 2004). Thus, while providing valuable insights that start to explain potential benefits of VR, literature in the field of advertising is lagging behind because companies are already exploring more advanced 3D ways, such as 360° advertising. And secondly, previous work about VR has mainly focused
on VR in an informational context used in other industries rather than advertising, such as education. However, at this moment, marketers in the field of advertising are leveraging the opportunities to break through the advertising clutter by using a more transformational context (Kerrebroeck et al., 2017). They are moving away from traditional informational ‘features-benefits’ advertising towards creating experiences for their consumers (Schmitt, 2000). Transformational advertisements are more creative in focussing on the experience consumers get while using a product, whereas informational advertisements focus on
elaborating about product attributes or benefits (Puto & Wells, 1984). These transformational advertisements can offer an even more advanced product experience with the help of VR. Considering the fact that there is not much research done about transformational VR in the field of advertising, previous work about informational VR in other industries is taken into account as preliminary research. Thus, currently there are two trends in advertising which induced the gap in advertising literature. Therefore, this study makes several contributions to the literature. First, building on the theory of 3D advertising, this study suggests that VR advertising has the possibility to generate the same positive behavioural outcomes amongst consumers by providing a more enhanced virtual product experience. And secondly, this study develops a new theory that suggests that these VR advertisements, that already proved to be successful in informational context in other industries, could also work in
transformational context in the advertising industry.
As such, the first goal of this study is to address the theoretical and managerial challenges by demonstrating that the use of VR in advertising positively influences different consumer behaviours, including purchase intention and perceived risk. Furthermore, while expecting a positive overall effect from the use of VR, this study investigates for which type of product (durable vs. convenience) its effectiveness is the greatest. And thirdly, to better understand the process behind the effectiveness of VR, the mediating role of perceived risk is
included to explain the effects. The research is conducted by means of an experiment to establish causal relationships. To reach these research goals, the following research question and sub-question are proposed:
RQ: “What is the effect of virtual reality use in advertisements on the purchase intention of
the advertised product amongst consumers?”
Sub-question: “To what extent is the effect of the virtual reality advertisement on purchase
intention mediated by the risk consumers perceive to make a purchase mistake?”
It is highly important for marketers in the digital landscape to obtain more knowledge about the VR approach with which they can influence their consumers most effectively. By learning about the process behind the effectiveness of VR as a marketing tool, behavioural patterns of consumers are better understood, which is extremely relevant for marketers. And since marketers are increasingly faced with the strategic decision to invest in various
technology platforms, these results could reassure them to adopt this technological innovation into their marketing campaigns successfully.
This study is structured as follows. First a theoretical framework and the hypothesized relationships will be presented which can be seen as the foundation of the study. Thereafter, the methodology is explained, including the experimental research design, data collection and measurement of the variables. This is followed by the findings of the study and the discussion of these results. Finally, a conclusion is drawn based on the complete study.
3. Literature Review
The emergence of VR technology heralds a new era for advertising. The potential of
visualizing advertisements in a 360° format has made VR evolve from a niche technology into the potential to reach the everyday consumer (Mollen & Wilson, 2010). As such, an
for marketers, since it provides insights to how their advertisements can be optimized to reach consumers effectively. As previously mentioned, VR in advertising is a relatively new
theoretical field since only a few innovative brands have started to explore its potential. Nevertheless, existing research has found that VR affects diverse marketing variables in other industries, such as attitude, perceived knowledge, recall and recognition (Yim, Cicchirillo, & Drumwright, 2012). However, these findings are mostly confined to the context of other industries with the focus on informational VR messages, leaving the potential effectiveness of transformational VR messages unknown. Within the advertising industry, prior research has already shown that using 3D technology in advertising has a great potential of influencing consumers as compared to regular 2D advertising. Therefore, it is important to identify if this also applies to the more advanced type of 3D advertising: VR advertising.
The potential of VR in advertising will be further explained in the following
theoretical framework that is needed to fill the identified literature gaps. First, the antecedents of VR effects are discussed to provide a better understanding of how the use of VR interfaces in advertising could potentially influence consumers positively. Thereafter, a discussion about the impact of VR effects in other industries is presented that might indicate if some of these effects are generalizable to advertising. Following, this study looks at what is already
researched about VR within the field of advertising. Since little is known about the potential of VR in this field, it is needed to explore which consumer behaviours reacted well to 3D advertising in previous research, a forerunner of VR. Finally, the framework presents the process that could be underlying the effects of VR on consumers to give marketers a better understanding of the behavioural patterns.
3.1 Antecedents of Virtual Reality Effects
learn about the antecedents in VR that could potentially stimulate certain consumer
behaviours. Steuer’s (1992) definition of virtual reality indicates that a virtual environment has the capacity to evoke the sensation of telepresence, which consists out of two factors:
vividness and interactivity. Telepresence is the extent to which someone feels present in a
mediated environment (Li, Daugherty, & Biocca, 2017). This means, the higher the level of telepresence in the advertisement, the stronger the level of virtual reality and its effects (Figure 1). This study theorizes if this could be established by enabling consumers to interact with products in virtual environments and letting them feel like they are actually experiencing the product.
Vividness and interactivity both partially enable consumers to have this feeling of telepresence. Vividness is defined as ‘the representational richness of a mediated
environment’ (Steuer, 1992). According to Kerrebroeck et al. (2017), the vividness aspect (i.e. the quality of the images and the sense of movement in the virtual environment) is the
dimension that generally has been emphasized in marketing so far. Vividness is used
interchangeably with the term media richness in prior research. The extent of media richness also depends on the level of representational quality and volume of content in the mediated environment (Suh & Lee, 2005). According to Suh and Lee (2005), VR provides a high media richness which allows consumers to examine realistic images of products from various angles and distances. Simultaneously, they found that VR makes consumers experience a strong feeling of telepresence, which engages them and increases their comprehension of the products (Suh & Lee, 2005).
Interactivity is defined as ‘the extent to which consumers can participate in modifying the form and content of a mediated environment’ (Kerrebroeck et al., 2017). The level of interactivity depends on the number of choices available in modifying the content of a medium (Choi, Miracle, & Biocca, 2001). In contrast to vividness, there is a scarce in VR
interfaces in marketing which provide interactivity (Kerrebroeck et al., 2017). As the interactivity in 360° advertisements is limited to only virtually modifying the angles, telepresence in this research will also mainly be manipulated through high levels of media vividness.
Telepresence is the key feature when it comes to advertisements that are designed for persuasion effects (Tussyadiah, Wang, Yung, & Dieck, 2018). According to Tussyadiah et al. (2018), this is because telepresence is a causal factor for human information processing performance. This means that telepresence could lead to a variety of effects such as
enjoyment and persuasion, which are primary goals of advertising (Lombard & Snyder-Duch, 2001). Choi, Miracle, and Biocca (2001) agree with this statement and explain that
telepresence could enhance a web user’s enjoyment, task performance, involvement, persuasion and memory, depending on the media content and the user characteristics. However, they also indicate that not much research is done regarding the effect of
telepresence on advertising effectiveness. Therefore, this study contributes to the theory of telepresence by researching if higher levels of interactivity and vividness in VR
advertisements actually lead to positive behavioural outcomes amongst consumers.
Figure 1 Antecedents of virtual reality effects
Vividness/Media Richness
Interactivity
3.2 Virtual Reality in other Industries
The visualization capabilities of VR have already proven to be a successful tool in health-related industries, the tourism sector, the gaming industry and education (Mikropoulos & Strouboulis, 2004; Tian et al., 2014; Tussyadiah et al., 2018). However, the question remains if these virtual reality effects are applicable to the field of advertising as well. Multiple
researchers in other industries have indicated that some of their results are useful predictors of the potential effectiveness of VR in marketing. This study further researches that potential. In the health-industry, the research of Tian et al. (2014) showed that the storytelling potential of VR in health-related informational videos creates a feeling of engagement and participation amongst patients in self-management of their diseases. Considering the fact that the use of VR storytelling in health-related industries builds a stronger experience in the understanding of the disease, they also suggest that the use of these virtual stories could build a stronger brand experience (Tian et al., 2014). Accordingly, this research investigates if the positive effects visible in the health industry are transferable to digital advertising strategies.
In addition to the above, research in training and education also proved that informational virtual reality leads to stronger learning outcomes amongst students
(Kerrebroeck et al., 2017). For example, the research of Mikropoulos and Strouboulis (2004) showed that a high level of telepresence in virtual educational environments causes high levels of cognitive performance and emotional development amongst pupils. These are both factors that contribute to knowledge construction. The VR part in this research concerned a game in which pupils interacted with an educational virtual environment to see if they could handle certain tasks (Mikropoulos & Strouboulis, 2014). The environmental richness and interactivity in the game created the high level of telepresence (Mikropoulos & Strouboulis, 2014). Marketers need to know if this level of telepresence in advertisements also has the possibility to contribute to the knowledge construction of brands in consumers’ minds. As
such, this study contributes to the academic literature by building on these findings and further investigating VR opportunities.
Beside these informational VR messages in the health industry and education, research has also shown the potential benefits of using transformational VR. For instance the research of Tussyadiah et al. (2018) about VR in tourism. They found that the sense of presence during a VR experience of a touristic destination could lead to positive consequences, including an increased level of preference in the touristic destination and visit intention. As such, they provide a theoretical explanation for the effectiveness of VR in influencing users’ response to marketing stimuli in the tourism sector. These results could be helpful for other marketers justifying investments in transformational VR, but they need to be further researched in this study.
Previous work in the gaming industry has also demonstrated the potential of transformational VR. Grigoroivi and Constantin (2004) researched the effectiveness of advertisements embedded in the virtual environment of gaming (e.g. driving a car of a certain brand). They claim that VR gaming increases users affective engagement with the
environment content due to their particular structural features such as high immersion. This, in turn, modifies the way embedded advertisements in games are processed. Their research showed a mixture of results in terms of advertising effectiveness by providing gamers with this virtual brand experience. For example, driving a BMW car in the VR game led to a high brand preference (Grigoroivi & Constantin, 2004). Therefore, Grigoroivi and Constantin (2004) recommend future research to look more into VR as an ‘experiential’ marketing channel that structurally has different effects compared to traditional media. They claim that this is due to the ability to let consumers experience the environment more and create some sort of arousal, which could impact advertising effectiveness. A such, the theorization of VR in experiential advertising in this study will provide a contribution to the existing literature.
The one thing that keeps reoccurring in each industry is that VR has the possibility to build a more immersive and rich experience which in turn leads to favourable behavioural outcomes. Considering the positive effects found in every industry, more research needs to be done about VR as a marketing tool in the advertising industry and the transferability of these effects. This research aims to fill the identified gap.
3.3 Virtual Reality in Advertising
As time progresses, advertising has always evolved along with new technology innovations (Kerrebroeck et al., 2017). Traditionally, advertising is defined as ‘a form of controlled communication that attempts to persuade consumers through the use of a variety of appeals and strategies, to use or buy a particular product or service’ (Lombard & Snyder-Duch, 2001). Although the central goal of advertising (i.e. to persuade consumers) has stayed the same over time, the media environment of advertising is changing, and as a result, the nature of
advertising is changing as well (Lombard & Snyder-Duch, 2001). Thus, advertising has already shown a big shift from traditional advertising (such as television commercials) to more interactive mobile advertising (such as online videos). Technologies such as 360° rotation videos and 3D product presentations belong to the new type of advertising and enable higher levels of vividness and interactivity as compared to traditional advertising (Choi & Taylor, 2014). However, new research is needed to explore the potential of these
technologies. According to Kerrebroeck et al. (2017), when comparing a VR or 3D
environment to a traditional 2D environment, the exposure to the VR advertisement creates a broader and deeper experience than the 2D advertisement. ‘Broader’ in the sense that VR is able to address multiple senses and ‘deeper’ in the sense that the quality of the environment is more realistic (Kerrebroeck et al., 2017). The effectiveness of these broader and deeper experiences in VR advertising are unknown and will be researched in this study.
More existing studies have already shown the benefits of 3D advertising in comparison to traditional 2D advertising. 3D advertisements can be seen as VR
advertisements in the beginning stages. This is because 3D advertising also refers to a virtual environment that uses high-end graphics to make consumers feel as if they physically
experience the product (Suh & Lee, 2005). For example, an experiment of Choi and Taylor (2014) looked at the advertising effectiveness of a 2D versus a 3D format for products such as watches and jackets. They found that the 3D format outperforms the 2D format in improving consumers’ purchase intention, attitude towards the brand, and intention to revisit the website. According to them, these 3D advertisements enhance the acquirement of visual sensory information, by for instance visualizing the product from different angles. This consequently drives the consumer to create a more vivid mental image of the product (Choi & Taylor, 2014).
Corresponding to Choi and Taylor (2014), most existing studies about VR in digital advertising look at 3D advertisements that are informational and allow consumers for instance to visualize the products in more advanced ways. For example, the research of Yim et al. (2012) exposed participants to KFC and Burger King food advertisements in a 3D format. In this 3D format, consumers were able to rotate the product images in the middle of the screen to provide a more detailed product visual. As such, they demonstrated that a 3D advertisement generates higher levels of vividness as opposed to flat advertisements, which consequently resulted in a positive attitude towards the advertisement.
However, more and more, marketers are shifting away from traditional informational ‘features-benefits’ advertising towards creating experiences for their consumers (Schmitt, 2000). Thus, in addition to informational advertisements, marketers have also started to explore the potential of VR in transformational advertisements (Kerrebroeck et al., 2017). These advertisements would fall under the bigger concept of experiential marketing which is
a form of advertising that focuses primarily on helping consumers to experience a product or service by making the advertisement more engaging (Tynan & McKechnie, 2010).
Williams (2006) indicates that in response to this growing trend of experiential marketing, VR technologies incorporating multimedia enable marketers to create memorable experiences for consumers that integrate meaning, consumption and brand loyalty. For example, existing research showed that driving a stimulated car model creates enjoyment and satisfaction which consequently leads to positive influences on the purchase intention for the actual car
(Papagiannidis, See-To, & Bourlakis, 2014). This suggests that a ‘hedonic’ virtual experience of the product can provide a foundation for favourable behavioural outcomes amongst
consumers. According to Papagiannidi et al. (2014), these results prove that experiencing products virtually plays an increasingly important role in the context of e-commerce. Nevertheless, as most existing studies in advertising focus on 3D or informational messages, the question remains if product experiences in transformational 360°
advertisements will also lead to positive consumer behaviour. This research attempts to answer that question. Theoretically, researching 360° advertisements provides a better
understanding of VR experiences that involve virtual depictions of using a product in the real world. Actions in these videos, such as sightseeing and navigation, resemble actual
consumption (Tussyadiah et al., 2018). Thus, this study will lead to a better conceptualization of the role of VR in constructing intentions towards actual consumption or purchase of the products.
In general, research about these 360° advertisements is limited due to the fact that the use of VR in advertising has not gone mainstream yet. A few innovative brands are already exploring and leveraging the potential of 360° advertisements. As a result, these
advertisements keep growing more valuable for generating engagement and delivering an impactful product experience for brands active in the digital landscape (Zhang, Cao, Ho, &
Chow, 2017). Some brands already use the VR tool in advertisements strategically to: 1. Communicate the brand’s mission through immersive storytelling (Tom’s)
2. Give assistance through a digital tutorial in VR for customers who want to learn certain skills with the offered products (Lowe’s)
3. Let the consumers virtually experience the features or functionality of the product (Mercedes) (Greenwald, 2016).
After looking further in the existing VR advertisements by some innovative brands, it appears that at this moment, VR is mainly used for immersive virtual product experiences. Consequently, this type of VR advertising will function as the independent variable in this research. Thus, since prior research demonstrates that 3D advertising and informational VR in other industries mainly have significant positive influences on individual behaviour, VR in advertising might also impact the consumer positively. The theoretical perspective that suggests that differences in the level of interactivity and vividness in advertising influence multiple important consumers behaviours is further developed in the following paragraphs. The hypothesized relationships are visualized in Figure 2.
3.4 Purchase Intention
Existing studies about VR in other industries substantiate its persuasive role by suggesting that the experience of telepresence in VR can translate into certain positive attitudes, inducing behavioural change, for instance in health behaviour (Fox, Arena, & Bailenson, 2009). The VR applications in advertising are also designed for various persuasive goals, such as the intention of consumers to purchase the advertised product. Nevertheless, little research is done about the potential of VR in advertising to persuasively increase the intention to purchase amongst consumers. Since prior research has shown that the purchase intention of consumers is the most common way to measure advertising effectiveness (Beerli & Santana,
1999), it is important to understand if VR has the potential to increase this intention. The intention to purchase a product is defined as ‘the likelihood that a customer will choose a certain brand of a product category in a specific buying situation’ (Crosno et al., 2009). The use of VR images in digital marketing communications has been gaining
prominence recently by enabling consumers to experience products realistically online (Suh & Lee, 2005). Suh and Lee (2005) already found that VR images on e-commerce channels increase the overall consumer learning (e.g. purchase intention). They explain that this is due to the possibilities of VR images to mitigate the problems associated with customers’ lack of physical contact with products. The question if these positive effects also apply to VR interfaces in videos needs to be further researched.
However, although it seems that VR has the potential to increase purchase intentions through offering an immersive virtual product experience, the rich environment could also be a distraction according to Nah, Eschenbrenner, and DeWester (2001). They found that
exposure to a 3D virtual environment generates both positive and negative effects on brand equity when compared to a 2D environment. The positive effects of the 3D virtual world on consumers occur through telepresence and enjoyment of the experience, which in turn affect the behavioural intention positively. The negative effects on consumers are explained by the distraction-conflict theory. This theory states that conflicts faced by consumers are a
consequence of highly interactive and rich media that result in distractions from attention to the brand (Nah et al., 2011). Therefore, they recommend marketers to take limitations in the attention span of consumers into account when designing such advertisements. According to them, marketers should do this to avoid cognitive overload which could lead to consumers being distracted from branding information (Nah et al., 2011). Since there are contrary results in the literature, it is needed to investigate if purchase intention actually is positively affected by VR in advertisements rather than 2D advertisements, or if VR does lead to a distraction.
Stated formally:
H1: There is a stronger positive relationship between the virtual reality advertisement and purchase intention than between the 2D advertisement and purchase intention.
3.5 Perceived Risk
Considering that previous research about 3D advertising has shown that VR has the potential to have positive consequences on consumer behaviour, this study attempts to learn more about the process behind the effectiveness of VR as a marketing tool. This could uncover the
behavioural patterns of consumers, which is extremely relevant for marketers. Knowing how to reduce the perceived risk of making a purchase mistake has become important in marketing studies since perceived risk has been regarded as a strong explanatory variable in consumer behaviour (Shim & Lee, 2011). Perceived risk refers to the subjective expectation of loss that may exist in the purchase of products or services (Taylor, 1974). Taylor (1974) suggests a risk reduction method in which consumers search for further information to reduce their
uncertainty. In other words, risk reduction can be determined by what kind of information consumers acquire. Previous studies found that although online advertisements are not able to provide tangible experiences for trying out products, the enhanced interactivity of 3D product images allow consumers to learn more about the product attributes (Shim & Lee, 2011). The enhanced interactivity is created by the possibility to see the product from all angles. This in turn, increases the amount of information acquired (Shim & Lee, 2011). As a consequence, this lowers the perceived risk and influences consumers’ behavioural intentions in the context of online advertisements according to Yang and Wu (2009). Although VR also enables consumers to acquire more information, no research has shown if VR has the same potential as 3D in minimizing these perceived risks. This study further examines this potential.
advertising derives its effectiveness from its ability to increase consumers’ confidence in their evaluation of the product. Their research is about the 3D visualisation of a product which enables consumers to virtually experience the product. The virtual experience in their research is defined as: ‘the conveyance of experiential product attributes in an online simulation of a direct experience’ (Debbabi et al., 2010). Klein (1998) explains how this virtual experience is able to reduce the risk prior purchase. He found that a virtual experience can ‘transform’ experience attributes (features assessed through actual use, e.g. taste and softness) into search attributes (features assessed without actual use of the product, e.g. size and colour). In a reaction to the study of Klein (1998), this study researches if the virtual experience in VR advertisements could actually transform experience attributes of all products into search attributes. That would mean that even the risk of buying a product such as tea could be reduced through the transformation of attributes by virtually experiencing the product. Assuming that 3D product experience is a forerunner of VR product experience, prior research indicates that VR should be able to minimize the risk of making a purchase mistake through increasing the information acquired about the product. However, previous research about VR did not take this variable into account as a mediator yet. The variable perceived risk could partially explain the positive relationship between VR advertisements and purchase intention. Therefore the following hypotheses are proposed:
H2: The positive relationship between the virtual reality advertisement and purchase intention is partially mediated by perceived risk.
H2a: There is a negative relationship between the virtual reality advertisement and perceived risk.
3.6 Product Experience
Given that a direct experience is generally the optimal way for consumers to learn more about the products before purchase, marketers could benefit from a simulated product experience online (Klein, 2003). Since online advertisements now have the capability to use the 360° VR technology, it is not overly ambitious to expect that the digital marketing landscape is able to provide those simulated product experiences. According to Suh and Lee (2005) VR enables consumers to learn about products more thoroughly because it provides them with three factors: interactivity with the product, an increased telepresence and high-quality three-dimensional angles of the product (Suh & Lee, 2005). However, existing research found that the effectiveness of these factors depend on the type of advertised product (Choi & Taylor, 2014). Since both brands with convenience and durable products are experimenting with VR in advertising, it is useful to understand the effectiveness for both type of products.
Choi and Taylor (2014) researched the moderating role of material versus geometric products on the relationship between 3D advertisements and the consumers' attitude towards the brand, their purchase intention, and intention to revisit the website. The difference between these products is that material products on the one hand are judged based on
properties such as texture, weight and temperature by touching them (Choi & Taylor, 2014). Whereas geometric products are judged based on properties such as size, appearance and shape by looking at them (Choi & Taylor, 2014). They found that 3D advertisements fail to have superior effects for material products (e.g. a watch) as compared to geometric products (e.g. a jacket). This can be explained due to the fact that consumers usually need real touch feedback before purchasing material products (Choi & Taylor, 2014). They claim that for geometric goods, the mental image induced by visually inspecting the 3D advertisement may evoke the illusion of trying out the product. The question if divergent products in VR
products remains and will be answered in this study.
However, the results in the research of Choi and Taylor (2014) could also be due to the fact that buying a watch is a higher involvement purchase than buying a jacket. Since a watch could be seen as a durable good which is not purchased as frequently and lasts longer than a convenience good such as a jacket. It could be the case that because these products differ in involvement, the perceived risks of purchasing the products also show a difference. Buying a durable high involvement product comes with much higher risks than with the purchase of low involvement products (Zeithaml, 1988). This is for example due to the monetary sacrifice, which is mostly higher for durable products (Zeithaml, 1988). When these purchase-risk perceptions are high, customers are highly inclined to search for product-related information (Bansal & Voyer, 2000). Therefore, the perceived risk is expected to be lower when consumers experience the product in a virtual environment before purchasing it (Shim & Lee, 2011). Because the perceived risk is already higher for durable goods, this study expects that by experiencing the durable product (a car) in a VR advertisement, the perceived risk of making a purchase mistake will be reduced more than for a convenience product (tea). Accordingly, this research will integrate the moderator product type (low involvement
convenience product vs. high involvement durable product) to see if consumers respond differently to these simulated product experiences. The following hypothesis is proposed:
H3: The positive relationship between the virtual reality advertisement and perceived risk is moderated by product type, so that this relationship is stronger for durable products than for convenience products.
3.7 Brand Evaluation
Brand evaluation refers to the subjective assessment of a brand by consumers as a result of prior investments in the marketing mix (Vogel, Evanschitzky, & Ramaseshan, 2008).
A primary determinant of the evaluative direction of information processing is someone’s prior evaluation of a brand (Vogel et al., 2008). Consumers who have a favourable prior brand evaluation are likely to be less critical of advertisements for that brand (Chattopadyay & Basu, 1990). Therefore, this variable will be measured and held statistically constant during the research to make sure that the effects are not due to cofounding factors, such as a prior favourable evaluation of the advertisement because of the brand.
Figure 2 Conceptual Model
4. Methodology
This chapter will first explain the research strategy and discuss the sample and the procedure of the experiment. Afterwards, the design and the manipulation in the experiment will be discussed with the measures and measurement scales of the variables.
VR Advertisement Perceived Risk Purchase Intention
Type of Product
H3
H1 (+)
H2
4.1 Research Strategy
To test the hypotheses, an experiment was conducted with a 2 (advertisement: VR vs. 2D) x 2 (product type: convenience vs. durable) full factorial between-subjects design. The research is a quantitative research where data was collected through an experiment embedded in a
survey. Using an online survey in Qualtrics made it possible to expose participants to the experimental intervention in the form of a manipulated online advertisement. As the main goal of this research was to assess the causal effect of VR use in advertisements on purchase intention, the research question was ideally suited for an experiment.
4.2 Sample and Procedure
Due to limited economic constraints for this experiment, a convenience sample was used through online surveys. The participants were selected based on convenient accessibility through social media and e-mail. The survey was distributed among a Dutch sample. After the distribution of the online surveys, 207 participants started the survey, of which 68% female and 32% male. The majority of the participants had a high education: WO (33%); HBO (29%) and were mainly below thirty years old (65%). After correcting for the participants who dropped out early, 198 participants were eventually taken into account for this study. The participants were randomly assigned to one of the four conditions and then exposed to a concise explanation of the research after which they filled in the informed consent form (Appendix B). They were assured of complete anonymity throughout the whole survey. In the second part, a pre-test was done in which the variables were measured when they were not influenced by the stimulus material yet. Subsequently, the participants were exposed to one of the four advertisements. Hereafter, the variables purchase intention and perceived risk were measured with the use of a seven-point Likert scale. All items of the different variables were randomly provided to each participant to prevent sequential effects.
Furthermore, the data collection was conducted according to the code of ethics of the University of Amsterdam.
4.3 Manipulation
4.3.1 Manipulation Check
Before the main study was conducted, a manipulation check was executed to ensure that the participants perceived the manipulation as intended. In this research, the level of telepresence in the advertisements was manipulated by exposing the participants to a VR advertisement versus a 2D advertisement. The level of telepresence was realized by increasing the levels of media vividness and interactivity in the experimental VR conditions. To make sure that the participants perceived the VR advertisements as more vivid and interactive than the 2D advertisements, six items were measured based on existing research. Vividness was measured on a seven-point Likert scale with statements from Kerrebroeck et al. (2017) such as ‘The advertisement makes it easy for me to imagine the use of the [product]’ and ‘The
advertisement makes it easy to picture myself riding/drinking the [product]’. In addition, interactivity was also measured on a seven-point Likert scale with the following statement from Song and Zinkhan (2008): ‘I feel that I had a great deal of control over the
advertisement’ (Appendix A for all items). Since vividness is a more important aspect of telepresence than interactivity according to Kerrebroek et al. (2017), this aspect was
emphasized in the manipulation check. The survey was distributed among 46 participants by a convenience sample.
4.3.2 Stimuli
To measure the effects of the VR advertisement in the main study, participants were randomly divided into two groups: the VR advertisement and the regular 2D advertisement. The level of
telepresence in the advertisements was manipulated through an increased level of interactivity and vividness in the VR advertisement as compared to the 2D advertisement. With this
method, the impact of the VR advertisement versus the regular 2D advertisement was examined, which both portrayed a product experience in a video representation. The experimental groups (the VR advertisement conditions) were exposed to a
YouTube video in 360° format. This enabled the participants to interact with the virtual world and modify the form and content of the mediated environment by using their mouse. Besides that, these conditions should evoke feelings of being present in the mediated environment and actually experiencing the product through the high level of vividness.
The control groups (the 2D advertisement conditions) were exposed to a similar YouTube video in 2D format which showed a regular advertisement of how it feels to
experience the product. Contrary to the experimental groups, these participants were exposed to a lower level of interactivity, considering they couldn’t modify the form and content of the mediated environment in 360 degrees. Besides, the level of vividness was also lower than in the experimental groups because they were exposed to an advertisement that focused less on the user perspective. This caused that it was harder to imagine actually experiencing the product.
However, to make sure that the videos in the control groups and experimental groups were still as similar as possible, all the advertisements were embedded in a transformational context which focused on the product experience and the feelings you get when using the product rather than focussing on the product attributes. Besides that, the type of product and length of the video were made similar as well to prevent cofounding factors as much as possible.
4.3.3 Product Type
The moderating effect of product type on the relationship between the VR advertisement and the perceived risk was measured for two type of products: a convenience product (Lipton Tea) versus a durable product (Mercedes car). Accordingly, this research measured if the effect of a VR advertisement reduces the perceived risk of making a purchase mistake more for a durable product than for a convenience product.
In the experimental groups, participants were exposed to a VR advertisement of the Mercedes car or the Lipton Tea. For the Mercedes condition, this means that the participants watched the advertisement from the perspective of a passenger in the car. This gave the participant a virtual experience of what it is like to drive the car. For the Lipton condition, the participant watched the advertisement from the perspective of someone who was drinking the tea. The idea here was to immerse the viewer in a cup of tea and to show the feeling you could get whilst drinking this tea. This was done with rich imagery and graphics that speak to the tea’s origin. In the control groups, participants were exposed to a regular non-interactive 2D advertisement of the same car or tea. These videos were also transformational in the sense that they focus on the experience and feelings you get while using the product. Table 1 provides an overview of the four conditions in this experiment.
Table 1 Online experiment: 2 X 2 full factorial between-subjects design
Variable Durable Product Convenience Product
VR advertisement VR – Durable Product (Car) VR – Convenience Product (Tea) 2D advertisement 2D – Durable Product (Car) 2D – Convenience Product (Tea)
4.4 Survey
All items used in the survey were derived from existing measurement scales to ensure a high construct validity. The full survey is presented in Appendix B. The measurement scales for each variable are highlighted below.
4.4.1 Perceived Risk
Perceived risk is considered a multidimensional concept involving performance risk and financial risk (Agarwal & Teas, 2001). Therefore, perceived risk was measured through scales used by Agarwal and Teas (2001) and Ha (2005) containing performance risk and financial risk items. The items were rephrased to make it possible that participants could answer on a seven-point Likert scale (1 = strongly disagree, 2 = disagree, 3 = somewhat disagree, 4 =
undecided, 5 = somewhat agree, 6 = agree, 7 = strongly agree). The three items for which the
participants had to indicate to what extent they agreed with the statements based on the advertisement were: ‘I think that the purchase of the advertised [product] will lead to a financial disappointment’, ‘I think that the [product] will/taste work satisfactorily’ and ‘I believe in the quality of the advertised [product]’ (Cronbach’s α = .92). The higher the score on this scale, the lower the perceived risk of making a purchase mistake.
4.4.2 Purchase Intention
The dependent variable, purchase intention, was measured as an unidimensional construct based on the existing scales of Azjen, Joyce, Sheikh, and Cote (2011), and Sweeney and Soutar (2001). The participants gave their answers on a seven-point Likert scale again (1 =
strongly disagree, 2 = disagree, 3 = somewhat disagree, 4 = undecided, 5 = somewhat agree,
6 = agree, 7 = strongly agree). The combined scale consisted out of four items on which participants had to answer the following statements: 'I want to buy the advertised [product] in
the future’, ‘I would not expect any problems with this [product]’, ‘I intend to buy the advertised [product]’ and ‘I would like to try the advertised [product]’ (Cronbach’s α = .91). The higher the score on this scale, the higher the intention of the participant to buy the advertised product.
4.4.3 Brand Evaluation
Brand evaluation was measured on a seven-point Likert scale adopted from the research of Dwivedi, McDonald, and Johnson (2014). The scale contained four items on which
participants had to respond to what extent they agreed with the following statements: ‘I would recommend [Brand] to friends’, ‘[Brand] is a likable brand’, ‘I would talk to other people about [Brand]’ and ‘[Brand] is a strong brand’ (Cronbach’s α = .84). Likewise, the
participants answered on a seven-point Likert scale (1 = strongly disagree, 2 = disagree, 3 =
somewhat disagree, 4 = undecided, 5 = somewhat agree, 6 = agree, 7 = strongly agree). The
higher the score on this scale, the more positively the brand was evaluated. This variable was used to exclude cofounding factors that affect the relationships in the conceptual model.
4.4.4 Control Variables
Besides brand evaluation, multiple variables were statistically held constant to test the relative impact of the VR product experience on the purchase intention correctly. Gender, age, and education were added as control variables because previous research for example showed that younger people are more adoption-ready to innovations such as VR (Rogers, 2010). Next to that, highly educated people have also proven to be more likely to positively adopt an innovation such as VR than lower educated people (Rogers, 2010). Age was captured in years, gender as male and female and education level as none, primary school, VMBO, HAVO/MAVO, VWO, MBO, HBO, WO, I don’t know or other.
5. Results
5.1 Manipulation Check
Before the experiment was conducted, the effectiveness of the manipulation was checked. Since literature showed that vividness and interactivity are the antecedents of virtual reality effects (Kerrebroeck et al., 2017), this check measured if the VR advertisement actually came across as more vivid and interactive than the 2D advertisement. With this check, the internal validity of the experiment was improved because it gave more certainty that the effects are due to changes in the construct that this study is interested in. Participants (N = 46) were randomly exposed to one of the four conditions and indicated to what extent they perceived the advertisement as vivid and interactive (Appendix A for the survey). First the VR
advertisement and the 2D advertisement of the durable product were tested. The VR
advertisement proved to generate stronger feelings of vividness and interactivity (Mv = 5.83,
SDv = .47; Mi = 5.88, SDi = .64) than the vividness and interactivity of the 2D advertisement (Mv = 2.49, SDv = 1.33; Mi = 2.15, SDi = 1.35). Likewise, the VR advertisement for the convenience product proved to generate stronger feelings of vividness and interactivity (Mv = 5.39, SDv = .68; Mi = 5.79, SDi = .89) than the 2D advertisement (Mv= 2.70, SDv = 1.71; Mi = 2.33, SDi = 1.75). The differences in vividness between the VR and 2D condition were significant for both the durable product (t(19) = -6.81, p = <.001) as for the convenience product (t(18) = -516, p = <.001). Furthermore, the differences in interactivity between the VR and 2D condition were also significant for both the durable product (t(19) = -7.28, p = <.001) and the convenience product (t(18) = -4.58, p = .004). This analysis showed that the manipulation of the independent variable in the stimulus material was done successfully (Appendix C for the stimulus material).
5.2 Hypothesis Testing
Prior to the hypothesis testing, a principal axis factoring analysis (PAF) and a bivariate analysis were conducted to compose the scales and to explore the correlations between them. The PAF was conducted on the following scales to assess measurement validity: brand evaluation1, perceived risk2 and purchase intention3. The Kaiser–Meyer–Olkin measure
verified the sampling adequacy for each variable in the analysis (KMO1 = .72; KMO2 = .70;
KMO3 = .80). Bartlett’s test of sphericity indicated that correlations between the items of each
variable were sufficiently large for PAF (χ² (6) = 75.13, p = <.001; χ² (3) = 123.35, p = <.001;
χ² (6) = 122.58, p = <.001). An initial analysis was run to obtain the eigenvalues for each
component in the data. For each variable, one component had an eigenvalue over Kaiser’s criterion of 1. These components explained 67.7% of the variance in brand evaluation, 86.6% of the variance in perceived risk and 78.6% of the variance in purchase intention. As such the results showed that all three factors were unidimensional constructs. In addition, the
composite reliability estimates for each scale were above the recommended .70, which indicates that there was an adequate consistency in the scales (Table 2 for Cronbach’s Alpha of each composed scale). The variables appeared to be normally distributed in the Skewness and Kurtosis test, because 0 fell within the calculated bounds of each 95% Confidence Interval.
The results above provided the evidence for the overall validity and reliability of the scales. Composite mean measures for brand evaluation, perceived risk and purchase intention were then constructed by aggregating the items. These measures were used in the following analyses. First a correlation matrix was conducted to explore the extent to which the
composed variables were related to each other (Table 2). As the Pearson correlation coefficients in this matrix show, the condition (VR advertisement vs. 2D advertisement) strongly correlates with both perceived risk (r = -.52, p = <.01) and purchase intention (r = .51, p = <.01). These results indicate that the variables are coherent with each other in the right direction. The higher the score on the advertisement (0 = 2D advertisement, 1 = VR advertisement), the lower the perceived risk and the higher the purchase intention. To further investigate the expectations, it was necessary to research if the condition actually is a
predictor of these variables. Table 2 Correlations, means and standard
deviations Variables M SD 1 2 3 4 5 6 7 8 1. Gender 1.68 .47 - 2. Age 32.6 13.1 -.13* - 3. Education 6.69 1.38 .11 -.41** - 4. Condition .48 .50 .05 -.07 -.04 - 5. Product Type .45 .50 .05 -.13* .02 0 - 6. Brand Evaluation 4.89 .98 -.04 .05 -.07 .01 -.03 (α = .84) 7. Perceived Risk 3.32 1.75 -.12 -.11 .09 -.52** -.01 -.25** (α = .92) 8. Purchase Intention 4.34 1.68 .02 .12* -.10 .51** .02 .29** -.85** (α = .91)
* Correlation is significant at the .05 level (2-tailed). ** Correlation is significant at the .01 level (2-tailed).
Subsequently, the hypothesis tests were conducted. Prior research suggests that age, gender and education influence an individual’s perception of innovations such as virtual reality (Rogers, 2010). Therefore, it was made sure that these variables were controlled for and held statistically constant in every analysis. The prior brand evaluation was also held constant to prevent cofounding effects. First, it was examined whether the type of
advertisement had an effect on the purchase intention of consumers as this was expected in H1. This hypothesis suggested that the participants exposed to a VR advertisement are more likely to have a high purchase intention of the advertised product in comparison to the participants exposed to a 2D advertisement of the same product. The analysis for H1 was done by means of a bivariate analysis: the t-test for differences between two averages. Since the Levene’s Test showed that there was a significant difference in the variance of purchase intention (F = 36.80, p = <.001) for the VR advertisement and the 2D advertisement, the t-test in which the equal variances were not assumed was used. The analysis of the results showed that there was a significant difference in purchase intention amongst consumers that were exposed to the VR advertisement as compared to the 2D advertisement (t (177) = -8.32, p = <.001, 95% CI [-2.10, -1.30]). Thereafter, the effect size of Pearson’s correlation coefficient was taken into account which actually showed how strong the relationship was between the two variables (Table 2). The correlation appeared to be strong according to Cohen’s rules of thumb for interpreting these effect sizes (r = .51, p = <.01) (Durlak, 2009). Participants who were exposed to the VR advertisement had on average a higher intention to purchase the advertised product (M = 5.22, SD = 1.13, N = 95) than participants who were exposed to the 2D advertisement (M = 3.52, SD = 1.70, N = 102) (Table 3). Therefore, H1 was supported.
Amber van der Neut (10565027)
In addition to the main effect found between the VR advertisement and purchase intention, H2 stated that this effect is partially mediated by an indirect effect of perceived risk. According to H2, it was expected that the perceived risk of making a purchase mistake is reduced by exposure to a VR advertisement of a product, which in turn explains the higher purchase intention. A multiple regression analysis was used to obtain the necessary
coefficients for the different components in the regression model in Figure 3. For this regression analysis, the PROCESS function in SPSS by Andrew F. Hayes was used to make sure that the variables age, gender, education and brand evaluation were controlled for. Further, the analysis of H2, H2a and H2b was divided in the following four steps to make the process clearer. Nevertheless, all the hypotheses (H2, H2a and H2b) were run in one go and the results were analysed from the same PROCESS model. The results of these hypotheses are summed up in Figure 3, Table 4 and Table 5.
In step 1 of this analysis the direct effect between the VR advertisement and perceived risk was calculated. In H2a, it was expected that exposure to a VR advertisement as compared to a 2D advertisement creates a lower perceived risk of making a purchase mistake. This relationship between the independent variable and the mediator is illustrated as path A1 in
Figure 3. The direct effect needed to be significant in order for the mediation effect to exist. As Figure 3 shows, the effect of the VR advertisement on perceived risk is A1 = -1.80. This
means that consumers that are exposed to a different advertisement (2D vs. VR) are estimated Table 3 Means and standard deviations of effect conditions on purchase intention
Condition M SD
VR Advertisement1 5.22 1.13
2D Advertisement2 3.52 1.70
to differ by respectively 1.80 units in perceived risk. Since the sign of A1 is negative, those
who are exposed to a VR advertisement are estimated to be lower in their perceived risk. This effect was statistically different from zero, t= -8.30, p = <.001, with a 95% confidence interval from -2.23 to -1.37. Therefore, H2a was supported.
In step 2, the direct effect between perceived risk and purchase intention was calculated. H2b stated the expectation that there exists a negative relationship between perceived risk and purchase intention. This means the lower the perceived risk, the higher the intention to purchase. This relationship between the mediator and the dependent variable is illustrated as path B1 in Figure 3. Similarly to the relationship A1, this relationship needed to
be significant in order for the mediation to work. The effect of the perceived risk on purchase intention appeared to be B1 = -.74. This means that those who are exposed to the same level
of the advertisement (e.g. both saw the VR advertisement), but differ by one unit in their level of perceived risk are estimated to differ by B1 = -.74 in purchase intention. The sign of B1 is
negative, meaning that those consumers relatively low in perceived risk are estimated to be higher in their purchase intention. The results showed that the perceived risk amongst consumers significantly predicts the purchase intention (t = -16.44, p = <.001, with a 95% confidence interval from -.83 to -.66). Accordingly, H2b was supported as well.
In step 3, the total effect and the direct effect between the VR advertisement and the purchase intention were calculated to reconfirm H1 and to see if there also exists a direct effect between the independent and dependent variable. The total effect and direct effect are illustrated as Path C1 and Path C2 in Figure 3. For the total effect of C1, the significant results
proved the positive effect of the VR advertisement on purchase intention which reconfirmed H1 (t = 8.36, p = <.001, 95% CI [1.32, 2.13]). Two consumers who differ by exposure to a VR advertisement vs. a 2D advertisement are estimated to differ by 1.73 units in their reported purchase intention. The positive sign means that the participants who saw the VR
advertisement reported to have a higher purchase intention of the advertised product than those who saw the 2D advertisement.
In addition, the results of the direct effect of C2 also proved that the relationship
between the VR advertisement and the purchase intention was statistically significant (t = 2.54, p = .012, 95% CI [0.09, 0.69]). The value of path C2 is .39, which means that those who
differ in exposure to a VR advertisement vs. a 2D advertisement and who experience the same level of perceived risk (i.e. who are equally confident in the purchase of a product), are
estimated to differ by .39 units in their purchase intention. This indicates that they are .39 units more prone to purchase the product if they have seen the VR advertisement as compared to the 2D advertisement, when their perceived risk is held constant. Moreover, the results in this step also showed that the overall model explains 74% of the variance in purchase intention, which was statistically significant as well (R2 = .74, p = <.001).
And lastly, step 4. In this step the indirect effect of the VR advertisement on purchase intention through the mediator perceived risk was measured. This step confirmed if the mediating effect really exists and therefore provided an answer to H2. The path of this relationship between the independent variable, mediator and dependent variable is illustrated as path A1 + B1 in Figure 3. The results showed that consumers who differ in exposure to the
two advertisements (VR vs. 2D) are estimated to differ by 1.34 units in their reported
purchase intention. This is a result of the tendency for those who feel less risk to purchase the advertised products, which in turn translates into a greater purchase intention. The indirect effect was statistically different from zero, as revealed by the 95% BC Bootstrap confidence interval that was entirely above zero (.94 to 1.74). Because both the direct effect and the indirect effect appeared to be significant, perceived risk is considered a partial mediator in this model. Consequently, the overall H2 was supported as well.
Table 4 Results regression analysis of effect advertisement type on perceived risk and
purchase intention
Table 5 Results regression analysis of direct effect advertisement type on purchase intention
and indirect effect on purchase intention through the mediator perceived risk
Coeff. SE p Path LLCI ULCI
Direct Effect 0.39 0.15 .012 C2 0.09 0.69
Total Effect 1.73 0.21 <.001 C1 1.32 2.13
Boot SE Path Boot LLCI Boot ULCI
Indirect Effect 1.34 0.20 A1+B1 0.94 1.74
Perceived Risk (M) Purchase Intention (Y)
Antecedent Coeff. SE p Path Coeff. SE p Path
VR advertisement (X) -1.80 0.22 <.001 A1 0.39 0.15 .012 C1
Perceived Risk (M) - - - - -0.74 0.05 <.001 B1
Constant 7.68 1.03 <.001 6.34 0.71 <.001
R2 = .34 R2 = .74
Figure 3 Unstandardized regression coefficients for the relationship between the
advertisement type and purchase intention, mediated by perceived risk. *p = <.05
Finally, in order to test H3, the moderated mediation effect of product type on the relationship between the VR advertisement and perceived risk was measured. The expectation of H3 was that the effect of the VR advertisement on perceived risk is stronger if the VR advertisement presents a durable good as compared to a convenience good. To analyse this hypothesis the scores of the independent variables were standardized to reduce
multicollinearity issues and to make the parameters comparable (Marquardt, 1980).
Thereafter, a multiple regression analysis was conducted. The results showed that 27% of the variance in perceived risk was explained by the product type, the type of advertisement and the interaction between the product type and the type of advertisement (R2 = .27). This proportion of explained variance is according to Kalmijn and Kraaykamp (1999) 1% above average compared to existing Dutch literature. However, the interaction effect of the product type and the type of advertisement on the perceived risk appeared to be quite low and not significant (t = .27, b* = .02, p = .786). Moreover, the main effect of the product type on the perceived risk did also appear not to be significant (t = -.15, b* = -.01, p = .885). The findings show that when the condition changes from a 2D to a VR advertisement, the difference in
Type of Advertisement Purchase Intention
C1 = 1.73*
C2 = .39* Perceived Risk
B1 = -.74*
perceived risk between participants who saw the convenience product or the durable product, increases by just .03 units. On the basis of this information it was possible to confirm that the product type does not have a moderating role or direct effect in this respect. The data from this regression analysis can be found in Table 6. Further, the means and standard deviations in Table 7 also show that the average perceived risk of the participants exposed to the
advertisements with a convenience good are very close to the average perceived risk of those exposed to the advertisements with a durable good. This proved that the moderation was not significant in the direction H3 expected.
.
Table 6 Results regression analysis of effects product type, advertisement type and product
type * advertisement type on perceived risk
Variable B SE b* t p
Constant 3.42 0.11 31.7 <.001
Zcondition (X) -0.90 0.11 -0.52 -8.35 <.001
Zproduct (M) -0.02 0.11 -0.01 -0.15 0.885
Interaction Product * Condition (XM) 0.03 0.11 0.02 0.27 0.786
Table 7 Means and standard deviations of the conditions split into two product type
dimensions on perceived risk
Condition Product Type M SD
VR Advertisement Durable Good1 2.50 0.14
Convenience Good2 2..47 0.16
2D Advertisement Durable Good3 4.24 0.28
Convenience Good4 4.33 0.24
