Sunteți pe pagina 1din 43

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/305714119

Can a virtual supermarket bring realism into the


lab? Comparing shopping behavior using virtual
and pictorial store...

Article in Appetite · July 2016


DOI: 10.1016/j.appet.2016.07.033

CITATIONS READS

7 332

4 authors, including:

Erica van Herpen Hans C. M. van Trijp


Wageningen University & Research Wageningen University & Research
60 PUBLICATIONS 626 CITATIONS 342 PUBLICATIONS 6,852 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Behavioural Science and the Law View project

Sustainable Packages View project

All content following this page was uploaded by Erica van Herpen on 02 August 2016.

The user has requested enhancement of the downloaded file.


Accepted Manuscript

Can a virtual supermarket bring realism into the lab? Comparing shopping behavior
using virtual and pictorial store representations to behavior in a physical store

Erica van Herpen, Eva van den Broek, Hans C.M. van Trijp, Tian Yu

PII: S0195-6663(16)30302-6
DOI: 10.1016/j.appet.2016.07.033
Reference: APPET 3090

To appear in: Appetite

Received Date: 2 February 2016


Revised Date: 14 June 2016
Accepted Date: 23 July 2016

Please cite this article as: van Herpen E., van den Broek E., van Trijp H.C.M. & Yu T., Can a virtual
supermarket bring realism into the lab? Comparing shopping behavior using virtual and pictorial store
representations to behavior in a physical store, Appetite (2016), doi: 10.1016/j.appet.2016.07.033.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT

Can a Virtual Supermarket Bring Realism into the Lab?

Comparing Shopping Behavior using Virtual and Pictorial Store Representations to

PT
Behavior in a Physical Store

RI
SC
Erica van Herpena*

Eva van den Broekb

U
Hans C. M. van Trijpc
AN
Tian Yud
M
D
TE

*
Corresponding author.
EP

a
Erica van Herpen is associate professor, Marketing and Consumer Behavior Group,
Wageningen University, Hollandseweg 1, 6706 KN Wageningen, The Netherlands (e-mail:
Erica.vanHerpen@wur.nl, Phone: +31 317 483385, Fax: +31 317 484361).
C

b
Eva van den Broek was a postdoc at Marketing and Consumer Behavior Group, Wageningen
AC

University, Hollandseweg 1, 6706 KN Wageningen, The Netherlands, at the time this


research was undertaken, and works at LEI Wageningen UR (e-mail: E.vandenBroek@wur.nl,
Phone: +31703358229).
c
Hans C. M. van Trijp is professor of Marketing and Consumer Behavior, Wageningen
University, Hollandseweg 1, 6706 KN Wageningen, The Netherlands (e-mail:
Hans.vanTrijp@wur.nl, Phone: +31 317 483385, Fax: +31 317 484361).
d
Yian Yu was employed by Noldus Information Technology, Wageningen, and Essensor,
Ede, at the time this research was undertaken.

1
ACCEPTED MANUSCRIPT
Abstract

Immersive virtual reality techniques present new opportunities for research into consumer

behavior. The current study examines whether the increased realism of a virtual store

compared to pictorial (2D) stimuli elicits consumer behavior that is more in line with behavior

in a physical store. We examine the number, variety, and type of products selected, amount of

PT
money spent, and responses to price promotions and shelf display, in three product categories

RI
(fruit & vegetables, milk, and biscuits). We find that virtual reality elicits behavior that is

more similar to behavior in the physical store compared to the picture condition for the

SC
number of products selected (Milk: Mstore = 1.19, Mvirtual = 1.53, Mpictures = 2.58) and amount

of money spent (Milk: Mstore = 1.27, Mvirtual = 1.53, Mpictures = 2.60 Euro), and for the selection

U
of products from different areas of the shelf, both vertically (purchases from top shelves, milk
AN
and biscuits: Pstore = 21.6 %, Pvirtual = 33.4 %, Ppictures = 50.0 %) and horizontally (purchase
M

from left shelf, biscuits: Pstore = 35.5 %, Pvirtual = 53.3 %, Ppictures = 66.7 %). This indicates that

virtual reality can improve realism in responses to shelf allocation. Virtual reality was not able
D

to diminish other differences between lab and physical store: participants bought more
TE

products and spent more money (for biscuits and fruit & vegetables), bought more national

brands, and responded more strongly to price promotions in both virtual reality and pictorial
EP

representations than in the physical store. Implications for the use of virtual reality in studies

of consumer food choice behavior as well as for future improvement of virtual reality
C
AC

techniques are discussed.

Keywords: virtual store; virtual reality; consumer; food; supermarket

2
ACCEPTED MANUSCRIPT
Acknowledgements

This work is part of the FOCOM project, which was supported by the European Regional

Development Fund and the Dutch provinces Gelderland and Overijssel (Grant number

2011P017004). The content of the paper reflects only the views of the authors. The authors

want to thank Leanne Loijens of Noldus Information Technology for her help with the initial

PT
study and Sylvia Kok of Essensor B.V. for her valuable input in setting up and collecting the

RI
data for the main experiment. We also thank Wim Vaessen of Essensor B.V. for stimulating

discussions and Ivo van der Lans for help in the analyses.

U SC
AN
M
D
TE
C EP
AC

3
ACCEPTED MANUSCRIPT
1. Introduction

The use of immersive virtual reality techniques to simulate a grocery store presents

exciting opportunities for research into consumer behavior. Through the use of virtual reality,

a 3D store environment can be simulated in a realistic and cost-efficient way (Berneburg,

2007; Bressoud, 2013; Ruppert, 2011). Data is collected in a tightly controlled but realistic

PT
store environment with high flexibility (Khan, Nuijten, and Deslé, 2011). In the context of

RI
food, virtual reality has been applied to study food cravings (Ledoux et al., 2013), the

emotional responses to food in patients with an eating disorder (Gorini et al., 2010), and as a

SC
screening tool for mild cognitive impairment (Zygouris et al., 2014). Moreover, virtual

supermarket systems have been used to study consumers’ reactions to price changes of food

U
products (Waterlander et al., 2011; 2012b; 2014), point-of-sale displays (Kim et al., 2014),
AN
emotional responses to retail environments (Massara et al., 2010), and responses to emptied
M

shelf space (van Herpen et al., 2009). A 3D virtual store has been called “an innovative and

unique research tool with great potential in the study of food choice behaviour” (Waterlander
D

et al., 2012a, p. 254) and the In-Store Marketing Institute predicts that the use of virtual
TE

reality in store simulation will become standard practice as an indispensable tool for

understanding in-store consumer behavior (Breen, 2009; Shankar et al., 2011).


EP

Virtual reality techniques are especially appealing because these can increase the

external validity of lab experiments. Traditionally, lab experiments form the basis of many of
C
AC

the empirical investigations in consumer research, but have been criticized for placing people

in artificial conditions where the link with real-life phenomena may be low (Winer, 1999;

Pham, 2013). The use of field studies has been promoted, but control over external variables

and practical issues can present problems for the execution of such field studies. Some

manipulations are impractical or costly (e.g., changing store layout), and there are often

external factors that may disrupt the study but cannot be controlled for (e.g., behavior of other

4
ACCEPTED MANUSCRIPT
shoppers). Another approach would then be to not bring the experiment to the real world, but

to bring the real world into the lab, that is, to ensure that the stimulus material and context

provided in the lab reproduces the actual decision-making situation (Levin, Louviere, and

Schepanski, 1983). The present study examines one way of doing so: using virtual reality to

better represent a food shopping environment in the lab.

PT
The key objective of the present study is to provide insight into the extent to which real-

RI
life shopping behavior in a supermarket can be captured in a lab environment when virtual

reality is used compared to a more traditional technique (i.e., pictorial representation of the

SC
shopping environment). It is important for both scholars and practitioners who contemplate

the use of virtual reality techniques for food shopping to understand which aspects of

U
consumer behavior can be better captured using these techniques, and for which aspects of
AN
consumer behavior the use of virtual reality compared to a pictorial representation does not
M

provide better insights (Campo et al., 1999). By comparing a choice task using virtual reality

to both a shopping trip in a real brick-and-mortar supermarket (with a similar choice task) and
D

a choice task using product pictures, we can examine to what extent the increased realism of
TE

the virtual store compared to pictorial stimuli will elicit consumer behavior that is more in line

with the behavior seen in an actual supermarket. Benchmarking against existing


EP

methodologies is important to gain insight into the relative strengths and weaknesses of these

methods (Burke, 1996), but this has often been overlooked in prior studies examining the
C
AC

validity of virtual reality (see Berneburg, 2007, for an exception).

1.1. Comparing virtual reality, pictures, and physical store

To assess the impact of virtual reality in a lab setting, a comparison needs to be drawn

with a relevant alternative. We will compare virtual reality to a 2D pictorial representation of

the same shopping environment, keeping the store assortment, display, and product

5
ACCEPTED MANUSCRIPT
information identical. The differences between virtual reality and pictures thus center on how

immersive the presentation of the store is, not the actual products or information that are

presented to people. This allows for a direct assessment of what 3D immersive virtual reality

offers above and beyond 2D pictorial representations.

These two laboratory conditions (virtual reality and pictures) are compared to a real-life

PT
physical store. As we are interested in the effects of virtual reality, and not in other factors

RI
that may induce differences between the lab and real life, we will control for such other

factors like sampling, instructions, assortment size, and shopping task. Thus, participants who

SC
visit the physical store will be drawn from the same panel as participants in the other

conditions, and will receive the same information about the study and the same instructions.

U
Furthermore, whereas other studies on virtual shopping have used smaller assortment sizes
AN
than present in reality (e.g., Campo et al., 1999), a factor that has been shown to influence
M

consumer behavior by itself (Iyengar & Lepper, 2000; Chernev, 2003), the current study will

employ the same product assortment in all conditions. This allows for a direct focus on the
D

effects of using virtual reality versus pictures versus a physical store, keeping other factors
TE

constant.

This approach extends prior research examining aggregate sales data (Campo et al.,
EP

1999) or actual supermarket purchases (Burke et al., 1992; Waterlander et al., 2015) as a

benchmark against shopping behavior in a virtual store. Differences between sales levels and
C
AC

virtual reality may potentially be caused by many factors, some of which can complicate a

direct comparison. Examples include the comparison of accumulated sales over multiple

shopping trips to choices made during a much shorter time interval in a virtual store, and

comparing sales from multiple different real-life stores to a single virtual store. In our study,

we will compare a single shopping trip in a virtual environment with a single shopping trip in

6
ACCEPTED MANUSCRIPT
the same real-life environment, to focus more directly on the effects of using virtual reality

per se.

1.2. Effects of virtual reality

Virtual reality increases the realism of the choice environment compared to lab setups

PT
using textual or pictorial stimuli. This added realism should increase feelings of being present

RI
in the virtual environment: participants feel as though they are in the store rather than the lab.

Such a feeling of presence has often been linked to virtual reality experiences, and is an

SC
important advantage of using virtual reality (Schuemie et al., 2001; Steuer, 1992; Witmer and

Singer, 1998; Lessiter et al., 2001). This feeling of presence could affect consumer behavior.

U
Table 1 summarizes the results of four prior studies that have examined the influence of
AN
virtual reality on consumer behavior. These studies differ from ours in the benchmark that
M

was taken, as indicated in the table. Moreover, the two studies from the 1990’s (Burke et al.,

1992; Campo et al, 1999) used a less sophisticated virtual environment simulation than what
D

is currently available, which may impact results. Given the scarcity of prior research in this
TE

domain, the hypotheses that we will develop are exploratory in nature. Based on prior

research, relevant indicators were selected to assess consumer behavior. Specifically, we will
EP

explore effects of the use of virtual reality on the number of products selected, the variety and

type of brands selected, the amount of money spent, and responses to in-store price
C
AC

promotions. We do not examine market shares of individual products because the purchase

rate is likely to be too low for any specific item to allow for meaningful comparisons (see also

Bressoud, 2013). In addition to the indicators based on prior research, we also examine

responses to display characteristics (e.g., purchase of products from eye-level), where we

expect that the ability of virtual reality to mimic the natural walking pattern of consumers

could induce more natural behavior. That is, whereas pictures provide consumers with a

7
ACCEPTED MANUSCRIPT
straight-on view of a shelf display, this is not how consumers typically see shelves when

walking through a store, and virtual reality is able to better imitate actual walking behavior.

As Underhill (1999, p. 78) puts it: “People face and walk forward. The implications of this are

enormous […]. It requires an effort to turn your head to one side or the other to see the

shelves or racks as you pass them”. Because of this effort, and due to the natural way in which

PT
people pay attention to their surroundings, consumers may respond differently to where

RI
products are located, depending on how the shelf is presented to them (straight on or while

walking past it). In the next paragraphs, we will review each of the indicators that we will

SC
examine in more detail.

_______________________

U
Insert Table 1 about here
AN
_______________________
M

Number of products selected. Prior investigations have reported a tendency for people to
D

select more products in lab experiments than in reality (according to sales data) (Burke et al.,
TE

1992). Partly this can be attributed to a lack of budget considerations, but fewer time and

space constraints (e.g., shopping cart, car) can also contribute. We would thus expect that
EP

people choose more products in laboratory conditions than in a physical store. Because the

use of virtual reality should draw people into the shopping experience more, and trigger their
C
AC

habitual purchase behavior, we expect this effect to diminish when virtual reality is used.

Level of variety seeking. The tendency of people to seek diversity in their choices of

products is called variety seeking (Kahn, 1995). Variety seeking can occur over the course of

multiple shopping trips, and prior research examined this (Campo et al., 1999; Burke et al.,

1992). Variety seeking is also reflected by people buying a varied set of products at one point

in time (Simonson, 1990), which is what we examine. For consumers, variety seeking is a

8
ACCEPTED MANUSCRIPT
means to obtain stimulation and alleviate boredom: it complicates the buying process and

ensures a “change of pace” (Steenkamp and Baumgartner, 1992). Given that in laboratory

conditions people are less distracted from the decision-making task (e.g., less cluttered

environment, fewer product categories, no other people), they may be inclined to examine and

choose more different options than in a physical store. This is in line with research showing

PT
that a less stimulating choice context increases the amount of variety seeking compared to a

RI
more stimulating choice context (Menon and Kahn, 1995). We would thus expect higher

levels of variety seeking in the lab.

SC
Type of brands selected. Laboratory settings can encourage more socially desirable

behavior, because people have a greater sense of being observed (Burke et al., 1992). This can

U
affect the types of brands that people buy, as it is more socially desirable to buy national
AN
brands than generics (Campo et al., 1999). Thus, as shown in the study of Campo and
M

colleagues (1999), people may buy fewer generics in lab conditions than in the physical store

and buy more national brands in lab conditions than in the physical store.
D

Amount of money spent. People tend to overestimate their preference for products in the
TE

hypothetical settings of lab experiments (List & Gallet, 2001). This type of effect also occurs

in the context of willingness-to-pay experiments (Blumenschein et al., 2008; Paradiso and


EP

Trisorio, 2001), where there are observed differences between hypothetical and real settings

such that people are willing to spend more money in the former than in the latter. In our
C
AC

situation as well, the amount of money spent could be higher in a lab experiment than in a

physical store where products are actually purchased. Given that people should exhibit more

habitual buying behaviors when virtual reality is used, compared to a less emerging pictorial

representation, we would expect these effects to be smaller in virtual reality conditions.

Response to in-store price promotions. Compared to real-life, prices and promotional

signs can be more salient in laboratory experiments (Nevin, 1974; Campo et al., 1999). In the

9
ACCEPTED MANUSCRIPT
laboratory, other visual and non-visual sensory input is more limited and participants may be

(much) more aware of price promotions. The simulated store environment is ‘cleaner’, less

cluttered with a lack of other people, shopping carts, and so on, so that sales signs should

draw relatively more attention. Additionally, the newness of the virtual reality technique may

lead participants to overly concentrate on their behavior (Berneburg, 2007). In combination

PT
with a cleaner and more sterile environment in laboratory settings, people may notice and

RI
respond to price promotions more readily, as Burke and colleagues have found in their study

(1992).

SC
Response to display characteristics. Many studies in brick-and-mortar supermarkets

have shown that people respond to the location in which products are placed. Products that are

U
placed at eye-level are sold more than products placed elsewhere on the shelf (Drèze, Hoch,
AN
and Purk, 1999; van Herpen, van Nierop and Sloot, 2012). This effect is likely to diminish
M

when products are shown in a picture format, as people can more easily see both the top

shelves (above eye height) and the bottom shelves (below eye height) than in a physical store.
D

Virtual reality could better represent purchases from different shelves, as we will explore.
TE

Table 2 provides an overview of our expectations.

_______________________
EP

Insert Table 2 about here

_______________________
C
AC

2. Initial study

A central premise to the present research is that 3D representation of products can add

realism to the choice environment by enhancing participants’ feeling of presence in the store

compared to the more traditional research approach of using 2D representations (pictures of

assortments). Before conducting the main experiment, we performed an initial study to test

10
ACCEPTED MANUSCRIPT
that assumption. A sense of “presence” is often considered key to virtual reality experiences

(Schuemie et al., 2001; Steuer, 1992; Witmer & Singer, 1998; Lessiter et al., 2001). It implies

that a person has a sense of being in the virtual environment (rather than being in the research

lab, or being absorbed in internal thought; Waterworth et al., 2010). Given that the objective

of this initial study is to get first insights into the differences between virtual reality and

PT
pictures, we used a mock shelf display as comparison rather than the actual physical store (as

RI
will be used in the main experiment).

Participants were 90 students (31 male, 59 female; age between 18 and 35 years). They

SC
were recruited by posters and flyers and randomly assigned to one of three groups: (a) a mock

shelf display, (b) a representation of this shelf display in virtual reality, and (c) a picture of the

U
same shelf display. The shelf display contained 16 different red wines with short descriptions
AN
and prices indicated, in addition to other products (fruit juices, teas, cereals, cookies). All
M

participants were above the national age restriction to buy alcohol. Wine was chosen as a

category in which participants were likely to spend some time making their choice, so as to
D

increase their exposure to the choice environment.


TE

The mock shelf display consisted of shelves that are comparable to those commonly

seen in supermarkets with shelf tags for the individual products. The display covered
EP

approximately 9 square meters. Participants could walk around freely and pick up products.

The virtual supermarket was displayed using a PC with three LCD screens of 42 inch each,
C
AC

which resulted in a 180 degree field-of-view. The shopping simulation software was

developed by Green Dino BV (www.greendino.nl). Participants could navigate through the

virtual supermarket using keyboard and mouse, and they could examine a product in more

detail by double clicking on it when standing in its vicinity. This provided them with an

enlarged front view of the product itself, and an enlarged view of the shelf tag. In the 2D

picture condition, participants first saw an overview picture of the store. Next, they viewed a

11
ACCEPTED MANUSCRIPT
picture of the wine shelf with numbered wine bottles, and they could select and view

individual pictures for each bottle. Pictures were presented on a 34 inch LCD screen.

After giving informed consent, participants in the virtual reality condition were given

the opportunity to practice with the equipment as long as they needed to maneuver and select

products comfortably. All participants were asked to buy one bottle of wine for a dinner with

PT
friends. After selecting a bottle of wine, they filled in a questionnaire, were thanked, and

RI
received a coupon for the university cafeteria of 2 Euro as a reward for participation. Presence

was measured by six items (see Appendix A, based on Witmer and Singer, 1998, α = .78).

SC
An analysis of variance showed a significant effect of condition on the level of presence

that participants reported (F(2, 87) = 16.43, p < .001). Planned contrasts showed that all three

U
means significantly differed from each other. Specifically, the perceived level of presence was
AN
higher in the virtual reality condition (M = 5.16) than in the picture condition (M = 4.38, p =
M

.001), but not as high as in the condition with the mock shelf display (M = 5.73, p = .018).

This implies that virtual reality is able to increase the feeling of presence compared to using
D

pictures of products.
TE

3. Main experiment
EP

Having established that virtual reality induces a higher level of presence than pictures, the

main experiment aims to extend this finding through a comparison with a real life
C
AC

supermarket situation, and an assessment of effects on consumer behavior measures.

3.1. Method

3.1.1. Participants and design

Participants were recruited by telephone from a consumer panel, one week before data

collection. Only regular buyers of the product categories involved (fruit/vegetables, cookies,

and milk) were selected. Another selection criterion was that participants needed to be regular

12
ACCEPTED MANUSCRIPT
visitors of the supermarket used in the study. Participants were randomly assigned to one of

three groups: (a) a real supermarket, (b) a virtual simulation of this supermarket, or (c) a

pictorial representation of this supermarket. The virtual simulation and pictorial representation

were designed as exact copies of the assortments that participants would find in the real

supermarket for the three product categories of interest.

PT
During recruitment, participants were asked to hold their purchase plans for fruit, milk

RI
and biscuits for several days, so they could (realistically) shop for these products during the

study. By asking participants to not purchase in the categories of interest for several days and

SC
to make these purchases during the study, we hoped to obtain realistic purchase behavior.

Data collection in the physical store took place one week prior to the data collection using

U
representations of this store, in order to allow for the pictorial and virtual reality
AN
representations to be matched closely to the real life situation (e.g., sales promotion material).
M

Participants received a voucher of 7.50 Euro as reward for participation. In the condition of

the real life store this voucher was directly redeemable for the products that they selected
D

during the study but could also be used for a future purchase. In the other (laboratory)
TE

conditions, participants were free to visit the store (located across the road) directly after the

study or at any later time to redeem the voucher. Vouchers were not set as a budget, and
EP

participants were free to choose products for any monetary amount that they wanted. During

data collection in the physical store, nine participants noted that they counted prices to arrive
C
AC

at the voucher value, seven participants bought products that were not part of the study, and

ten participants did not fill in the questionnaire correctly. These participants were disregarded.

This left a total of 100 participants in the final dataset (68 female and 32 male participants;

age ranged from 18 to 74 years, with an average of 42 years).

13
ACCEPTED MANUSCRIPT
3.1.2. Stimuli

To examine a broad range of consumer behavior, three product categories were selected

of which the real-life supermarket assortment was completely rebuilt in 3D (virtual

supermarket condition) or 2D (pictorial representation). Specifically, we wanted to focus on

habitual purchases, purchases in which sensory aspects are important, and impulse purchases.

PT
First, as a high rate of repeat purchasing is indicative of more habitual buying (Adamowicz

RI
and Swait, 2012), a product category with a high level of repeat purchases was chosen. This

was the milk category. Second, the product category of fruit and vegetables was chosen

SC
because sensory aspects are especially important in this category. People often squeeze or

smell these product before purchasing (Burke et al., 1992). Third, an extensive product

U
category in which impulse buying is not uncommon was chosen: the biscuit category. A priori
AN
we do not expect differences in the ability of virtual reality to increase realism to the choice
M

task and affect purchase behavior measures across the three product categories. Rather, we

consider these product categories as replications, and use them to examine the generalizability
D

of our results.
TE

Data collection in the physical store condition took place in a local supermarket. This

supermarket contained 31 milk products (none of which had a promotional offering during the
EP

data collection period), 32 fruit/vegetable products (five of which were on sale during data

collection), and 230 biscuits (nine of which were on sale). The virtual supermarket showed a
C
AC

realistic simulation of the physical store, as the screen shots in Appendix B illustrate. All

products from the three target product categories (milk, fruit/vegetables, and biscuits) were

presented, in the same shelf position as in the physical supermarket, and participants could

walk around in the virtual environment using a predefined shopping route and scrutinize

products by using the keyboard and mouse. Sales promotions were visible through the same

promotional devices (shelf tags, posters) as were used in the physical store, in realistic size

14
ACCEPTED MANUSCRIPT
and the same in-store position. In the virtual store, the cookies took up one side of an aisle,

facing the fruit and vegetable department, and the milk products were placed in another aisle.

This second aisle also contained products from other categories, as indicated in the layout of

Appendix C, to provide context and increase realism.

The picture condition showed static screenshots of the virtual store on a regular desktop

PT
screen (19 inch). Screenshots showed (a) a view of the store, (b) an overall view of the target

RI
aisles, (c) a closer view of the shelves with index numbers assigned to the products, and (d)

pictures of the each individual product as well as its shelf tag. These were connected using

SC
hyperlinks, so that participants could switch from product to product, or from shelf to shelf.

3.1.3. Procedure
U
AN
In the physical store condition, participants were welcomed by a research assistant at the
M

entrance of the supermarket. They received a brief task instruction and a shopping list

containing the three focal product categories. During the shopping trip, they were
D

unobtrusively observed by two other research assistants at the fruit/vegetables, biscuit and
TE

milk aisles. The research assistants did not make themselves known to the participant,

pretended to be other shoppers, and noted their observations out of sight of the participant.
EP

After completing their shopping task, participants would meet a fourth research assistant at

the cashier desk and they received vouchers for their purchases. They also filled in a
C
AC

questionnaire at a table near the exit of the supermarket

The procedure in the other two conditions was as similar as possible, with the exception

that participants in the virtual reality condition could first practice in a different virtual store

for as long as they needed to feel comfortable with the setup. Participants received the same

instruction and shopping list. The virtual reality software recorded which products were

bought and in which amount, except for the fruit and vegetable products where participants

15
ACCEPTED MANUSCRIPT
manually wrote down the amounts they wished to buy. In the picture condition, participants

wrote down their choices for all product categories. After finishing their shopping trips, all

participants filled in the same questionnaire and were rewarded with store vouchers.

3.1.4. Measures

PT
Through either observation or automatic recording, the type and number of products

RI
chosen were recorded, and product prices were noted. Table 3 provides the measurement

details for the dependent variables of interest. To assess variety in the chosen cookies, the

SC
products from the assortment were categorized according to type (6 levels, e.g., biscuit, cake,

multilayer) and flavor (14 levels, e.g., chocolate, almond, cream). Variety across the chosen

U
cookies was assessed by calculating the entropy in the chosen set for these two attributes (see
AN
van Herpen & Pieters, 2002) and averaging across these.
M

Additionally, perceived presence was included in the questionnaire (see Appendix A for

details). To extend beyond the results of the initial study and also examine potential
D

subdimensions of presence, we used a more extended version of the presence scale (based on
TE

the advanced Witmer-Singer PQ; Witmer et al., 2005). Items were worded in the first person

and measured on 7-point scales with endpoints labelled “strongly disagree” to “strongly
EP

agree”. A principle component analysis using oblimin rotation pointed towards three

underlying factors, as indicated in Appendix A, which we labelled “involvement/control”,


C
AC

“immersion”, and “interface quality” (see also Witmer and Singer 1998 and Schuemie et al.,

2001, for a discussion on subdimensions of presence).

_______________________

Insert Table 3 about here

_______________________

16
ACCEPTED MANUSCRIPT
Background variables were product category experience and demographics. Product

category experience was assessed by asking how often in general the participant purchased

food products from the target categories on a 7-point scale ranging from daily to almost never.

The final questions contained demographic variables including age, gender, family

composition as well as experience in shopping in the specific supermarket of interest.

PT
Participants also noted for how many days they were shopping, the extent to which they

RI
choose the products that they usually buy (overall and for each of the three product categories

separately, 7-point scale), and their general frequency of purchasing in the categories of

SC
interest.

3.2. Results
U
AN
3.2.1. Demographics and other sample characteristics
M

Participants were selected to be regular visitors of the supermarket that was used in the

study, and 54% of participants visited this supermarket at least once a week, whereas only
D

12% visited it less than once a month. Most participants were living together/married with
TE

children living at home (42%), living together/married without children (19%) or single

without children (15%). They shopped for 1 to 4 people, with an average of 2.2 persons for
EP

which they were buying. Almost all participants indicated that they shopped for consumption

by themselves or family members (95%). When asked for how many days they were
C
AC

shopping, most participants answered that they shopped for the next 2 to 7 days (92% of

participants answered within this range for fruit and vegetables, 94% for milk, and 67% for

biscuits). Furthermore, most participants indicated that they generally bought at least once a

week from the category (97% for fruit and vegetables, 95% for milk, and 75% for biscuits).

Participants also answered questions on whether they bought the products they usually

buy for each of the product categories. A repeated measures ANOVA with product category

17
ACCEPTED MANUSCRIPT
as the within subject factor and condition as the between subject factor showed no significant

effect of condition (F(2, 97) = 1.25, p = .291) and no significant interaction effect (F(4, 194)

= 0.87, p = .484). Across conditions, participants generally indicated that they bought the

products that they would usually buy. There was a significant effect of product category (F(2,

194) = 15.79, p < .001), indicating that buying what one usually buys was more prominent for

PT
milk (M = 6.04) than for either fruit (M = 5.65, p = .027) or biscuits (M = 5.11, p < .001),

RI
while fruit and biscuits did not significantly differ.

Despite random assignment, family composition was not completely balanced across

SC
conditions. In the physical store, more people were present who were married / living together

with children living at home than in the other conditions (67.6 % versus 13.3 % for virtual

U
supermarket and 39.4 % for picture condition). This is also reflected in the number of people
AN
participants say they are buying for. In the physical store condition this is significantly higher
M

(M = 2.46) than in the virtual supermarket (M = 1.93; p = .01) and marginally higher than

picture condition (M = 2.09, p = .06), while the latter two do not significantly differ. This
D

implies that one would expect people in the physical store condition to buy slightly more
TE

products than in the other conditions based on family composition, which is opposite to our

expectations. Other demographics did not significantly differ between conditions.


EP

3.2.2. Feelings of presence


C
AC

An ANOVA showed that the effects of condition on the level of presence did not reach

significance (F(2, 97) = 1.89, p = .16), although means are in the expected direction.

Examining each of the three subdimensions separately showed that “involvement/control” and

“interface quality” did not significantly differ between conditions (F(2, 99) = 0.80, p = .453

and F(2, 99) = 2.26, p = .110, respectively). The subdimension “immersion” differed between

conditions (F(2, 99) = 4.10, p = .020): in the virtual store participants indicated higher levels

18
ACCEPTED MANUSCRIPT
of being immersed in the shopping (M = 5.20) than in either the picture condition (M = 4.60; p

= .025) or the physical store (M = 4.50, p = .008), while the latter two conditions did not

significantly differ from each other. Thus, even though the virtual reality was able to enhance

presence in the (virtual) shopping environment compared to the picture condition, in this case

it exceeded the reported levels in the physical store when it comes to being completely

PT
immersed in the shopping process.

RI
3.2.3. In-store behavior

SC
An overview of results for the various measures related to in-store behavior is provided in

Table 4.

U
_______________________
AN
Insert Table 4 about here
M

_______________________
D

Number of products selected. The number of products selected in each condition was
TE

assessed using ANOVA, as indicated in Table 4. Because the distribution of this variable was

somewhat skewed, with many participants buying one or two products, especially in the milk
EP

category, we also analyzed the data using a non-parametric test (Kruskall-Wallis) and using

bootstrapping. Results were robust, and interpretation remained the same regardless of
C
AC

analysis method.

For both fruit and vegetables and biscuits, the number of products selected was higher in

the laboratory conditions than in the physical store. Because for fruits, the total number of

products selected requests the summing across products of very different types and sizes, we

also examined the number of different types of fruit/vegetable products selected. Results were

similar: more different types were selected in both virtual and picture conditions than in the

19
ACCEPTED MANUSCRIPT
physical store condition. In contrast to our expectations, virtual reality was unable to

significantly diminish the increased purchase amounts in lab conditions. Yet, results for milk

were in line with our expectations. For milk, participants were also inclined to buy more milk

products in the picture condition than in the physical store. Virtual reality diminished this

effect: in the virtual store, participants choose significantly fewer milk products than in the

PT
picture condition, and the number of products chosen did not significantly differ between

RI
virtual and physical store conditions. Here, thus, the virtual store better resembled the buying

behavior of the physical store than the picture condition did.

SC
Level of variety seeking. For biscuits, we examined the level of variety in the set of

products that participants chose. To assess variety seeking tendencies, we categorized this

U
vast category (over 200 different biscuits) into different types and flavor of biscuits, and
AN
examined the extent to which the sets that participants chose exhibited variety across these
M

types and flavors. Because many participants (31 %) chose only one packet of biscuits

(implying no variety seeking), the level of variety seeking was correlated highly with the
D

number of products that people selected (r = .82). Therefore, we examined variety seeking
TE

only for those participants who bought more than one packet of biscuits. As indicated in Table

4, participants chose a more varied set of biscuits in both the picture and virtual reality
EP

condition than in the physical store. In other words, participants seemed more inclined to seek

variety under lab conditions than in the real store.


C
AC

Type of brands selected. For both milk and biscuits, we examined whether participants

bought at least one generic / store branded product and whether participants bought at least

one product with a national brand. For the purchase of generic / store brands, we found no

significant effects. Purchase of national brands also did not differ across conditions for milk.

The number of people buying biscuits from national brands, however, did differ across

20
ACCEPTED MANUSCRIPT
conditions (χ2(2) = 11.39, p = .003; see Table 4). In line with our expectations, in the physical

store, fewer people bought national brands than in the picture condition and the virtual store.

Amount of money spent. Given that the number of products bought differs across

conditions, it is unsurprising that the total amount of money spent shows a similar pattern of

results. For both fruits & vegetables and biscuits, results showed that participants spent more

PT
money in both lab conditions than in the physical store. For milk, results are slightly different

RI
and in line with the number of products bought, showing that the amount of money spent

differed between the picture condition and the physical store, but not between the virtual

SC
reality condition and the physical store. Virtual reality was thus able to bring spending to the

level of the actual store for this category. For all three categories, the average price paid per

U
product did not differ across conditions (F(2, 97) = 1.41, p = .250 for fruit & vegetables; F(2,
AN
97) = 0.14, p = .867 for milk; F(2, 97) = 0.11, p = .896 for biscuits).
M

Response to price promotions. During the time of the study, price promotions were

present for the fruit & vegetables and the biscuits categories. For biscuits, only 12% of
D

participants bought one or more biscuits on sale, and this percentage did not significantly
TE

differ across conditions. For fruit & vegetables, purchase of products on price promotion was

more prominent. The number of participants who bought fruit on sale was high in both the
EP

virtual store condition (93.3 %) and the picture condition (90.9 %), and significantly higher

than in the physical store condition (62.2 %). This supports the expectation that in-store sales
C
AC

promotions are more salient in lab conditions than in the physical store, and thus lead to

stronger responses.

Response to display characteristics. For both milk and biscuits we examined the extent to

which people bought from top shelves (above eye-level), middle shelves, and bottom shelves.

Only for top shelves did we find differences. For milk, products from the top shelf were rarely

chosen in the physical store (2.7% of participants), but more often in the picture condition

21
ACCEPTED MANUSCRIPT
(33.3%, χ2(1) = 11.52 p = .001) and also more often, albeit to a lesser extent, in the virtual

store condition (16.7%, χ2(1) = 3.96, p = .047). For biscuits, the pattern was similar. In line

with expectations, a lower percentage of participants bought from the top shelf in the physical

store (40.5%) than in the picture condition (66.7%; χ2(1) = 4.78; p = .029), whereas the virtual

store was in-between and did not significantly differ from either (50.0%).

PT
For the biscuit category, we were able to also investigate horizontal display effects, that

RI
is, the extent to which participants took products from the left, middle, or right side of the

display. Results showed marginal differences for the left shelf (χ2(2) = 5.86; p = .053). Further

SC
examination showed that a lower percentage of participants bought from the left shelf in the

physical store (35.1%) than in the picture condition (63.6%; χ2(1) = 5.67; p = .017), whereas

U
the virtual store was in-between and did not significantly differ from either (53.3%). Thus,
AN
whereas the picture condition led participants to buy more from the left of the display than
M

they seemed to do in a real store, this effect was no longer apparent in the condition with

virtual reality.
D
TE

4. General discussion

The key objective of the current study is to provide insight into the extent to which, for
EP

research purposes, real-life shopping behavior can be captured in the lab to a greater extent by

the help of virtual reality techniques, as compared to more traditional approaches with
C
AC

pictures. Across the measures that were examined, two types of result patterns appear. On the

one hand, there are measures for which virtual reality improves the representation of the real

store in the lab (picture and physical store conditions differ, whereas virtual reality condition

and physical store condition do not significantly differ), and on the other hand, there are

measures for which the differences between real store and lab persist regardless of the use of

22
ACCEPTED MANUSCRIPT
virtual reality (both picture and virtual reality differ from physical store, whereas picture and

virtual reality do not significantly differ). We will discuss both groups of measures in turn.

Virtual reality was shown to better represent the behavior in the physical store than the

picture condition for the number of products selected and amount of money spent (for milk

only), and for the selection of products from different areas of the shelf, both vertically and

PT
horizontally. This shows that virtual reality indeed brings out more realistic purchase amounts

RI
in the milk category, a category in which habits are relied upon to a great extent. Moreover,

virtual reality improves responses to shelf allocation: it better represents choice behavior in

SC
the real store when it comes to the extent to which products from top shelves and from the left

side of the display are chosen. This is likely due to the way the display is presented. When a

U
product display is shown as a picture, participants view the display only as if they were
AN
standing in front of the display looking directly at it. This is not the way they are confronted
M

with products in an actual store. There, they often first view a product display at an angle,

walk towards it, and approach it from the side (as when walking through an aisle). This is
D

mimicked in the virtual reality but not by presenting pictures.


TE

For other measures, virtual reality and pictures both differed from the physical store.

Participants bought more products and spent more money (for biscuits and fruit &
EP

vegetables), exhibited more variety seeking, bought more national brands, and responded

more strongly to price promotions. Given that assortment size and presentation, as well as the
C
AC

size, layout, and format of the promotion material, was similar across conditions, these effects

may be due to people feeling less restricted in their purchases (e.g., no budget restrictions)

and being less distracted from the buying task (e.g., no other shoppers). This implies that

when the objective of a study is to assess purchase amounts, market shares of national brands,

or the size of responses to price promotions, virtual reality may not be very helpful in

obtaining better estimates than other lab-based approaches using pictures. This is not to say

23
ACCEPTED MANUSCRIPT
that virtual reality cannot be helpful in obtaining insightful results when it comes to

comparing the effects of different types of price promotions or in-store displays (see Kim et

al., 2014 and Waterlander et al. 2014 for examples). When comparing different types of

promotions, virtual reality may very well be used and should be able to help identify which

promotion is likely to lead to the largest consumer response. What our results suggest is that,

PT
although the identification of the most effective promotion may be unaffected, the size of the

RI
response is likely to be an overestimation of the actual consumer response in the physical

store (see also Bressoud et al., 2013).

SC
Another finding from our study was that the level of perceived immersion was higher in

the virtual than in the physical environment. This result is not in line with the classical

U
definition of presence as the “perceptual illusion of non-mediation” (Lombard and Ditton,
AN
1997), in which presence in virtual reality cannot exceed presence in real life. However, it is
M

in line with a more recent conceptualization of presence as a feeling of being located in the

perceived external world around the self, rather than being absorbed in the internal world of
D

thought (Waterworth et al., 2010), as well as with empirical findings that presence can be
TE

higher in virtual reality than in reality (Villani et al., 2012).

Our results have important implications for practitioners and scholars who contemplate
EP

the use of virtual reality in their study of in-store buying behavior. Whether virtual reality is

advisable clearly depends on the research question of interest. Especially when effects of
C
AC

product location in a display are crucial for the study at hand, the use of virtual reality can

provide good insights. Virtual reality may also be particularly suited to examine new products

at early stages in their development, in a close-to-market setting, to assess which packaging or

assortment display may be most effective. When using virtual reality for such applications,

our study suggests that absolute market shares may not be particularly well estimated, and this

24
ACCEPTED MANUSCRIPT
should be taken into account. Comparisons across different options (for packaging, display,

sales promotions, etc.) may provide better insights than absolute estimates of market shares.

Our results also have important implications for the further development of virtual reality

as a tool for the study of consumer buying behavior. For instance, participants have a

tendency to buy more products in lab condition than in a physical store. Future improvements

PT
in the virtual reality setup could attempt to diminish this difference. For instance, participants

RI
may be better able to track their purchases when the virtual reality setup is able to better

represent the movement of placing products in a shopping cart, to show the contents of the

SC
virtual shopping cart as people move along, to let participants carry a basket or push a

(virtual) shopping cart which increases in weight as more products are bought, and/or to have

U
participants go through a realistic (virtual) payment procedure. This may help reduce the
AN
difference in the number of products bought that we found, as future research could
M

investigate. In other aspects, a higher degree of realism through virtual reality may be

obtainable in the future as well, for instance when virtual reality techniques may incorporate
D

control devices that better resemble realistic action (e.g., grabbing of products), textures of
TE

products can be sensed through touch, and the social environment of other shoppers is

simulated (for a more thorough discussion of such aspects, see Tal and Wansink, 2001).
EP

Future research could also assess whether imposing budget restrictions in virtual reality

(as done in prior research, e.g., Waterlander et al., 2011) may induce participants to buy less
C
AC

and lead to purchase amounts that are closer to what is observed in an actual store. As recent

research indicates that buying with a budget in mind differs in important ways from buying

without a budget (van Ittersum et al., 2013), assessing the impact of budget restrictions on

various purchasing behavior indicators would be a promising avenue for future research.

Virtual reality applications could be used to investigate such effects of budget restrictions.

25
ACCEPTED MANUSCRIPT
The current study examined consumer behavior for three products categories, which

differ in the extent to which purchases are habitual, the importance of sensory aspects, and the

amount of impulse purchasing. To the extent that results patterns are similar across these

categories, we would expect that the insights generalize to other categories of fast moving

consumer goods. Future empirical testing is needed, however, to examine this in more detail.

PT
Future empirical testing is also needed in larger samples, to investigate whether more subtle

RI
differences may exist in moving from photographs to virtual reality, which were not picked up

in our study due to our relatively small sample size. Additionally, our study compared

SC
behavior in virtual reality to behavior in a physical store, while keeping other aspects

(shopping task, recruitment criteria, etc.) constant. This allowed us to focus on the specific

U
effects of virtual reality. Still, future research could also compare against naturally occurring
AN
behavior among customers who are unaware of being part of a study, to further generalize the
M

results.

In our research, we have used a virtual reality system in which participants are seated
D

behind a computer and see the virtual environment on large screens. This type of system
TE

differs from reality in multiple aspects, including a lack of physical movement and the use of

computer-based equipment for maneuvering. As technology advances rapidly at the moment,


EP

alternatives are becoming available in which some of these aspects can become closer to

reality. For instance, the use of virtual reality headsets and new interaction devices can
C
AC

encourage more natural movements in the virtual environment, which could stimulate more

realistic consumer behavior, and the current development of such devices is thus promising.

Other differences in the application of virtual reality may likewise affect the realism of the

behavior that is seen. For instance, previous virtual reality studies have used only a subset of

the assortments that are present in real life (e.g., Campo et al., 1999). As assortment size has

been shown to affect purchase behavior and the decision process (Iyengar and Lepper, 2000),

26
ACCEPTED MANUSCRIPT
this could potentially decrease the correspondence between behavior in reality and virtual

reality.

Future research should also examine the concept of presence more thoroughly. Prior

studies provide different accounts on the concept itself and the subdimensions of which it may

consist (Schuemie et al., 2001; Witmer et al, 2005). Our results indicate that presence may be

PT
a relevant concept, and the initial study has shown that virtual reality can increase the level of

RI
presence. Yet, in the main experiment the level of immersion (one of the subdimensions of

presence that we identified) was higher in the virtual store than in the physical store. This is

SC
likely due to people feeling more immersed in the shopping process itself and less distracted

by other aspects of the store (e.g., other shoppers, or thought or memories that are irrelevant

U
to the shopping process). Still, for the conceptualization of presence it seems potentially
AN
troublesome that people report a higher level of presence in a virtual than in a real
M

environment, and more research into the concept itself and its likely subdimensions seems

warranted.
D

In summary, the virtual store has a lot to offer to practitioners and scholars who want to
TE

gain more insights into in-store consumer behavior, using controlled lab conditions. Our study

provides initial evidence for advantages of using virtual reality rather than photographs in
EP

consumer behavior research. Specifically, our results indicate that virtual reality can stimulate

more habitual buying processes and ensure that responses to display characteristics (i.e.,
C
AC

where products are located) are in line with actual behavior in a real store. Still, people are

inclined to buy more in general, to buy more varied products, to buy more national brands,

and to buy more products with a sales promotion when in the lab than in the real store, and

this needs to be taken into account when interpreting findings from a study using virtual

reality.

27
ACCEPTED MANUSCRIPT
References

Adamowicz, W. L., & Swait, J. D. (2012). Are food choices really habitual? Integrating

habits, variety-seeking and compensatory choice in a utility-maximizing framework.

American Journal of Agricultural Economics, 1-25.

Berneburg, A. (2007). Interactive 3D simulations in measuring consumer preferences: Friend

PT
or foe to test results. Journal of Interactive Advertising, 8(1), 1-37.

RI
Blumenschein, K., Blomquist, G. C., Johannesson, M., Horn, N., & Freeman, P. (2008).

Eliciting willingness to pay without bias: Evidence from a field experiment. The

SC
Economic Journal, 118, 114-137.

Breen, P. (2009). Shaping retail: The use of virtual store simulations in marketing research

U
and beyond. In-Store Marketing Institute white paper,
AN
http://kelley.iu.edu/cerr/files/09ismi_virtualretailing.pdf
M

Bressoud, E. (2013). Testing FMCG innovations: Experimental real store versus virtual.

Journal of Product & Brand Management, 22(4), 286-292.


D

Burke, R. R., Harlam, B. A., Kahn, B. E., & Lodish, L. M. (1992). Comparing dynamic
TE

consumer choice in real and computer-simulated environments. Journal of Consumer

Research, 71-82.
EP

Burke, R. R. (1996). Virtual shopping: Breakthrough in marketing research. The Journal of

Product Innovation Management, 6(13), 558-559.


C
AC

Campo, K., Gijsbrechts, E., & Guerra, F. (1999). Computer simulated shopping experiments

for analyzing dynamic purchasing patterns: validation and guidelines. Journal of

Empirical Generalisations in Marketing Science, 4(2), 22-61.

Chernev, A. (2003). When more is less and less is more: The role of ideal point availability

and assortment in consumer choice. Journal of Consumer Research, 30(2), 170-183.

28
ACCEPTED MANUSCRIPT
Drèze, X., Hoch, S. J., & Purk, M. E. (1994). Shelf management and space elasticity. Journal

of Retailing, 70(4), 301-326.

Gorini, A., Griez, E., Petrova, A., & Riva, G. (2010). Assessment of the emotional responses

produced by exposure to real food, virtual food and photographs of food in patients

affected by eating disorders. Annals of General Psychiatry, 9, 30-39.

PT
Van Herpen, E., & Pieters, R. (2002). The variety of an assortment: An extension to the

RI
attribute-based approach. Marketing Science, 21(3), 331-341.

van Herpen, E., Pieters, R., & Zeelenberg, M. (2009). When demand accelerates demand:

SC
Trailing the bandwagon. Journal of Consumer Psychology, 19(3), 302-312.

van Herpen, E., van Nierop, E., & Sloot, L. (2012). The relationship between in-store

U
marketing and observed sales for organic versus fair trade products. Marketing Letters,
AN
23, 293-308.
M

van Ittersum, K., Wansink, B., Pennings, J. M. E., & Sheehan, D. (2013). Smart shopping

carts: How real-time feedback influences spending. Journal of Marketing, 77(6), 21-36.
D

Iyengar, S. S., & Lepper, M. R. (2000). When choice is demotivating: Can one desire too
TE

much of a good thing? Journal of Personality and Social Psychology, 79(6), 995-1006.

Kahn, B. E. (1995). Consumer variety-seeking among goods and services: An integrative


EP

review. Journal of Retailing and Consumer Services, 2(3), 139-148.

Khan, V.-J., Nuijten, K., & Deslé. (2011). Pervasive application evaluation within virtual
C
AC

environments. PECCS, 261-264.

Kim, A. E., Nonnemaker, J. M., Loomis, B. R., Shafer, P. R., Shaikh, A., Hill, E., Hooloway,

J. W., & Farrelly, M. C. (2014). Influence of point-of-sale tobacco displays and graphic

health warning signs on adults: Evidence from a virtual store experimental study.

American Journal of Public Health, 104(5), 888-895.

29
ACCEPTED MANUSCRIPT
Lessiter, J., Freeman, J., Keogh, E., & Davidoff, J. (2001). A cross-media presence

questionnaire: The ITC-Sense of Presence Inventory. Presence: Teleoperators and

Virtual Environments, 10(3), 282-297.

Ledoux, T., Nguyen, A. S., Bakos-Block, C., & Bordnick, P. (2013). Using virtual reality to

study food cravings. Appetite, 71, 396-402.

PT
Levin, I. P., Louviere, J. J., & Schepanski, A. A. (1983). External validity tests of laboratory

RI
studies of information integration. Organizational Behavior and Human Performance,

31, 173-193.

SC
List, J. A. & Gallet, C. A. (2001). What experimental protocol influence disparities between

actual and hypothetical stated values? Environmental and Resource Economics, 20, 241-

254.
U
AN
Lombard, M., & Ditton, T. (1997). At the heart of it all: The concept of presence. Journal of
M

Computer‐Mediated Communication, 3(2), 0-0.

Massara, F., Liu, S. S., & Melara, R. D. (2010). Adapting to a retail environment: Modeling
D

consumer–environment interactions. Journal of Business Research, 63(7), 673-681.


TE

Menon, S., & Kahn, B. E. (1995). The impact of context on variety seeking in product

choices. Journal of Consumer Research, 22, 285-295.


EP

Nevin, J. R. (1974). Laboratory experiments for estimating consumer demand: A validation

study. Journal of Marketing Research, 11(3), 261-268.


C
AC

Paradiso, M. & Trisorio, A. (2001). The effect of knowledge on the disparity between

hypothetical and real willingness to pay. Applied Economics, 33, 1359-1364.

Pham, M. T. (2013). The seven sins of consumer psychology. Journal of Consumer

Psychology, 23(4), 411-423.

Ruppert, B. (2011). New directions in the use of virtual reality for food shopping: marketing

and education perspectives. Journal of Diabetes Science and Technology, 5(2), 315-318.

30
ACCEPTED MANUSCRIPT
Schuemie, M. J., Van Der Straaten, P., Krijn, M., & Van Der Mast, C. A. (2001). Research on

presence in virtual reality: A survey. CyberPsychology & Behavior, 4(2), 183-201.

Shankar, V., Inman, J. J., Mantrala, M., Kelley, E., & Rizley, R. (2011). Innovations in

shopper marketing: Current insights and future research issues. Journal of Retailing,

87S(1), S29-S42.

PT
Simonson, I. (1990). The effect of purchase quantity and timing on variety-seeking behavior.

RI
Journal of Marketing Research, 27, 150-162.

Steenkamp, J. B. E., & Baumgartner, H. (1992). The role of optimum stimulation level in

SC
exploratory consumer behavior. Journal of Consumer Research, 434-448.

Steuer, J. (1992). Defining virtual reality: Dimensions determining telepresence. Journal of

Communication, 42(4), 73-93.


U
AN
Tal, A. & Wansink, B. C. (2011). Turning virtual reality into reality: A checklist to ensure
M

virtual reality studies of eating behavior and physical activity parallel the real world.

Journal of Diabetics Science and Technology, 5(2), 239-244.


D

Underhill, P. (1999). Why we Buy: The Science of Shopping. New York, NY: Touchstone
TE

Villani, D., Repetto, C., Cipresso, P., & Riva, G. (2012). May I experience more presence in

doing the same thing in virtual reality than in reality? An snwer from a simulated job
EP

inverview. Interacting with Computers, 24, 265-272.

Waterlander, W., Scarpa, M., Lentz, D., & Steenhuis, I. (2011). The virtual supermarket: An
C
AC

innovative research tool to study consumer food purchasing behaviour. BMC public

health, 11(1), 589-599.

Waterlander, W., Mhurchu, C. N., & Steenhuis, I. (2012a). The use of virtual reality in

studying complex interventions in our every-day food environment. Virtual Reality –

Human Computer Interaction. Ed. Xinxing, T., INTECH Open Access Publisher. doi:

10.5772/46410.

31
ACCEPTED MANUSCRIPT
Waterlander, W. E., Steenhuis, I. H. M., de Boer, M. R., Schuit, A. J., & Seidell, J. C.

(2012b). The effects of a 25% discount on fruits and vegetables: results of a randomized

trial in a three-dimensional web-based supermarket. Journal of Behavioral Nutrition and

Physical Activity, 9, 11-22.

Waterlander, W. E., Mhurchu, C. N., & Steenhuis, I. H. M. (2014). Effects of a price increase

PT
on purchases of sugar sweetened beverages. Results from a randomized controlled trial.

RI
Appetite, 78, 32-39.

Waterlander, W. E., Jiang, Y., Steenhuis, I. H. M., & Mhurchu, C. N. (2015). Using a 3D

SC
virtual supermarket to measure food purchase behavior: A validation study. Journal of

Medical Internet Research, 17(4), e107.

U
Waterworth, J. A., Waterworth, E. L., Mantovani, F., & Riva, G. (2010). On feeling (the)
AN
present. Journal of Consciousness Studies, 17(1-2), 167-188.

Winer, R. S. (1999). Experimentation in the 21st century: The importance of external validity.
M

Journal of the Academy of Marketing Science, 27(3), 349-358.


D

Witmer, B. G., & Singer, M. J. (1998). Measuring presence in virtual environments: A


TE

presence questionnaire. Presence, 7(3), 225-240.

Witmer, B. G., Jerome, C. J., & Singer, M. J. (2005). The factor structure of the presence
EP

questionnaire. Presence, 14(3), 298-312.

Zygouris, S., Giakoumis, D., Votis, K., Doumploulakis, S., Ntovas, K., Segkouli, S.,
C
AC

Karagiannidis, C., Tzovaras, D., & Tsolaki, M. (2014). Can a virtual reality cognitive

training application fulfill a dual role? Using the virtual supermarket cognitive training

application as a screening tool for mild cognitive impairment. Journal of Alzheimer’s

Disease, 40, 1-10.

32
ACCEPTED MANUSCRIPT
Appendix A: Items of the presence scale

Initial study

1. I was able to completely survey and search the environment using vision.
2. I completely engaged my visual senses when choosing the wine.
3. I had the feeling that I could move freely in the shopping environment.

PT
4. I was able to concentrate on choosing the wine and was not distracted by any devices
used during the task.
5. I felt involved in the selection of the wine.

RI
6. I was able to closely examine the wine bottles.

SC
Main experiment

First subdimension: involvement/control


1. I was able to search the shopping area completely by looking around.

U
2. I was able to take full control of the events that occurred while shopping.
3. I felt involved in the shopping trip.
AN
4. I had all my senses fully engaged in the shopping trip.

Second subdimension: immersion


M

5. I was completely unaware of events that took place outside the shopping area.
6. There were moments when I felt completely focused on doing the shopping.
7. There were moments when I felt completely focused on the retail environment.
D

Third subdimension: interface quality


8. I felt I could walk around freely in the store.
TE

9. I was able to examine the products closely.


10. I was able to concentrate on my purchase decisions.
11. I found it easy to move from shelf to shelf.
C EP
AC

33
ACCEPTED MANUSCRIPT
Appendix B: Screen shots and pictures of the main experiment

Setup of the virtual reality display (left) and picture display (right)

PT
RI
SC
The fruit and vegetables shelves in reality (left) and virtual reality (right)

U
AN
M
D
TE
C EP
AC

34
ACCEPTED MANUSCRIPT
The milk shelf in reality (left) and virtual reality (right)

PT
RI
U SC
AN
The biscuits aisle in reality (left) and virtual reality (right)
M
D
TE
C EP
AC

35
ACCEPTED MANUSCRIPT
Appendix C: Layout of the virtual environment (top view)

PT
RI
U SC
AN
M
D
TE
C EP
AC

36
ACCEPTED MANUSCRIPT
Table 1. Summary of prior literature

Study Comparison Dependent variables Main outcomes

Burke et al., Rudimentary Market share Less market share differences in soft
1992 lab (verbal drink category
descriptions) Response to price Stronger in rudimentary lab than real
vs. realistic
promotions supermarket
lab (VR) vs.

PT
supermarket Number of products Higher in both lab conditions than
purchase selected real supermarket
data Purchase of store brands Mixed results

RI
Brand switching Less switching in both lab
conditions than real supermarket

SC
Number of brands in the Smaller number of brands in both
choice set lab conditions than real supermarket

Campo et VR vs. Number of products Higher in VR than scanner data

U
al., 1999 scanner data purchased
AN
Purchase of national Higher in VR than scanner data
brands
Purchase of generics Lower in VR than scanner data
M

Purchase of private No significant difference


labels
D

Variety seeking No significant difference


Response to sales Indication for stronger responses in
TE

promotions VR than scanner data

Bressoud, VR vs. Attitude towards Less favorable in VR than in real


2013 experimental innovative product store
EP

real store Purchase intention No significant difference


Time spent purchasing a Higher in VR than in real store
C

product innovation
Purchase rate of an No significant difference, probably
AC

innovation due to low purchase rates overall


Waterlander VR vs. Proportional Similar overall purchase patterns
et al., 2015 supermarket expenditures on food
purchase groups
data
Number of products Similar overall purchase patterns
purchased
Total absolute Lower in VR than real supermarket
expenditures

Note. VR = Virtual Reality

37
ACCEPTED MANUSCRIPT
Table 2. Expectations

Variable Virtual store compared to Pictures compared to

physical store physical store

Number of products selected + ++

Level of variety seeking + +

PT
Purchase of store brands / generics - -

RI
Purchase of national brands + +

Amount of money spent + ++

SC
Response to price promotions + +

Selection of products at eye-level - --

U
AN
M
D
TE
C EP
AC

38
ACCEPTED MANUSCRIPT
Table 3. Measurement details

Variable Measurement

Number of products selected Count of the number of products chosen. For fruit &
vegetables, since product sizes differ greatly, we also
inspected the number of different items1 bought.

Level of variety seeking Assessed for biscuits. The 229 items in the category

PT
were content-coded according to two attributes: type of
cookie and flavour. Type of cookie was coded into 6
levels: cake, multi-layered, sprits (a typical Dutch type

RI
of cookie), biscuit, child-oriented, or other. Flavour was
coded into 14 levels: butter, nuts, almond, fruit (with
exception of apple), apple, chocolate, coffee, sugar,

SC
cream, speculoos, fibres, caramel, coconut, or other. For
each participant, the proportion of chosen products with
each attribute level (pl) was calculated, and the entropy
per attribute, calculated as ∑L –pl(ln(pl)). This was

U
averaged across both attributes (type and flavour) to
obtain a measure for the level of variety seeking.
AN
Purchase of store brands / generics Yes/no: is a product with a store brand or a generic
selected (not relevant for fruit & vegetables)
M

Purchase of national brand Yes/no: is a product with a national brand selected (not
relevant for fruit & vegetables)
D

Amount of money spent Total price of selected products


TE

Response to price promotions Yes/no: is a promoted product selected

Purchase from top / middle / Yes/no: is a product from this shelf selected (not
bottom shelves relevant for fruit & vegetables, due to the use of low
EP

shelves).

Purchase from left / middle / right Yes/no: is a product from left / middle / right shelf
C

shelves selected (only for biscuits; not relevant for milk due to
the small width of the display area).
AC

1
Measured as stock-keeping-units (SKU’s).

39
ACCEPTED MANUSCRIPT
Table 4. Summary of results

Variable Physical Virtual Pictures F-value χ2-value p-value


store reality (n = 33)
(n = 37) (n = 30)

Number of products selected


Fruit & vegetables 1.57a 3.50b 3.33b 20.60 <.001
(0.77) (2.00) (1.27)

PT
Milk 1.19a 1.53a 2.58b 17.04 <.001
(0.46) (0.82) (1.52)
Biscuits 1.70a 2.50b 3.03b 8.54 <.001
(0.81) (1.43) (1.72)

RI
Number of types (SKUs) selected
Fruit & vegetables 1.43a 3.40b 3.09b 23.94 <.001
(0.73) (1.79) (1.18)

SC
Level of variety seeking
Biscuits, without people who buy .52a .74b .78b 3.72 .029
just one product (0.32) (0.35) (0.33) (df = 2, 66)

U
Purchase of store brand
Milk (% buying) 78.4 % 66.7 % 75.8 % 1.26 .533
AN
Biscuits (% buying) 81.1 % 86.7 % 90.9 % 1.42 .493
Purchase of national brand
Milk (% buying) 21.6 % 40.0 % 33.3 % 1.26 .254
Biscuits (% buying) 27.0 % 63.3 % 60.6 % 11.39 .003
M

Amount of money spent


Fruit & vegetables 3.30a 7.05b 7.42b 17 <.001
D

(1.79) (4.55) (3.13) .20


Milk 1.27a 1.53a 2.60b 14.37 <.001
(0.78) (0.94) (1.44)
TE

Biscuits 2.01a 3.00b 3.67b 7.93 <.001


(0.89) (1.82) (2.32)
Response to price promotions
EP

Fruit & vegetables (% buying) 62.2 % 93.3 % 90.9 % 13.60 .001


Biscuits (% buying) 16.2 % 13.3 % 6.0 % 1.78 .412
Purchase from top/middle/bottom shelves
C

Milk (% buying top) 2.7 % 16.7 % 33.3 % 11.60 .003


Biscuits (% buying top) 40.5 % 50.0 % 66.7 % 4.84 .089
AC

Milk (% buying middle) 37.8 % 53.3 % 51.5 % 2.00 .368


Biscuits (% buying middle) 64.9 % 66.7 % 66.7 % 0.03 .983
Milk (% buying bottom) 62.2 % 50.0 % 54.5 % 1.04 .595
Biscuits (% buying bottom) 29.7 % 46.7 % 51.5 % 3.78 .151
Purchase from left/middle/right shelf
Biscuits (% buying left) 35.5 % 53.3 % 63.6 % 5.86 .053
Biscuits (% buying middle) 51.4 % 63.3 % 66.7 % 1.90 .386
Biscuits (% buying right) 37.8 % 50.0 % 54.5 % 2.11 .348

40
ACCEPTED MANUSCRIPT
Note. Unless otherwise indicated, means are provided for each condition with standard deviations

between brackets. Means with different superscripts (indicate with the letters a and b) are significantly

different based on LSD posthoc tests. Unless otherwise indicated, degrees of freedom for the ANOVA

tests reported in the table are 2, 97. Degrees of freedom for the Chi-square tests reported are 2.

PT
RI
U SC
AN
M
D
TE
C EP
AC

41

View publication stats

S-ar putea să vă placă și