Content uploaded by Qiuzhen Wang
Author content
All content in this area was uploaded by Qiuzhen Wang on May 04, 2016
Content may be subject to copyright.
This article was downloaded by: [Zhejiang University]
On: 13 July 2014, At: 02:10
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Enterprise Information Systems
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/teis20
The effect of human image in B2C
website design: an eye-tracking study
Qiuzhen Wang a , Yi Yang b , Qi Wang c & Qingguo Ma a
a School of Management , Zhejiang University , Hangzhou , China
b Fuzhou University Zhicheng College , Fuzhou , China
c NetEase, Inc. , Hangzhou , China
Published online: 13 Jun 2014.
To cite this article: Qiuzhen Wang , Yi Yang , Qi Wang & Qingguo Ma (2014) The effect of human
image in B2C website design: an eye-tracking study, Enterprise Information Systems, 8:5, 582-605,
DOI: 10.1080/17517575.2014.925585
To link to this article: http://dx.doi.org/10.1080/17517575.2014.925585
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or
howsoever caused arising directly or indirectly in connection with, in relation to or arising
out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
The effect of human image in B2C website design: an
eye-tracking study
Qiuzhen Wang
a
, Yi Yang
b
, Qi Wang
c
and Qingguo Ma
a
*
a
School of Management, Zhejiang University, Hangzhou, China;
b
Fuzhou University Zhicheng
College, Fuzhou, China;
c
NetEase, Inc., Hangzhou, China
(Received 31 December 2013; accepted 14 May 2014)
On B2C shopping websites, effective visual designs can bring about consumers’
positive emotional experience. From this perspective, this article developed a research
model to explore the impact of human image as a visual element on consumers’online
shopping emotions and subsequent attitudes towards websites. This study conducted
an eye-tracking experiment to collect both eye movement data and questionnaire data
to test the research model. Questionnaire data analysis showed that product pictures
combined with human image induced positive emotions among participants, thus
promoting their attitudes towards online shopping websites. Specifically, product
pictures with human image first produced higher levels of image appeal and perceived
social presence, thus stimulating higher levels of enjoyment and subsequent positive
attitudes towards the websites. Moreover, a moderating effect of product type was
demonstrated on the relationship between the presence of human image and the level
of image appeal. Specifically, human image significantly increased the level of image
appeal when integrated in entertainment product pictures while this relationship was
not significant in terms of utilitarian products. Eye-tracking data analysis further
supported these results and provided plausible explanations. The presence of human
image significantly increased the pupil size of participants regardless of product types.
For entertainment products, participants paid more attention to product pictures inte-
grated with human image whereas for utilitarian products more attention was paid to
functional information of products than to product pictures no matter whether or not
integrated with human image.
Keywords: human image; emotion; attitude towards online shopping websites; image
appeal; social presence; eye tracking
1. Introduction
In today’s Internet economy, a well-designed Business-to-Consumer (B2C) website has
the capacity to be a strategic competitive tool for e-retailers and thus become an important
research topic in the information systems (IS) area. Prior research has largely focused on
the usability of websites. The main dimensions of usability are effectiveness, efficiency
and satisfaction (De Angeli et al. 2003). Existing literatures mainly focus on users’
cognitive processes and behaviours during website browsing process with the aim to
explore how websites’effectiveness and efficiency have been improved by technology,
while less attention has been paid to online users’emotional experiences elicited from
multi-sensory elements (Cyr et al. 2009).
*Corresponding author. Email: maqingguo3669@zju.edu.cn
Enterprise Information Systems, 2014
Vol. 8, No. 5, 582–605, http://dx.doi.org/10.1080/17517575.2014.925585
© 2014 Taylor & Francis
Downloaded by [Zhejiang University] at 02:10 13 July 2014
However, emotional responses may determine which websites people choose to use as
they seek pleasure or enjoyment beyond task efficiency (Bucy 2000). Positive emotional
experiences will promote a web user’s behavioural tendency to approach the website
(Deng and Poole 2010). Prior research has revealed that this tendency will lead to longer
website browsing time, deeper exploration of products, more intense responses to promo-
tional incentives and enhanced probability of making a purchase (Menon and Kahn 2002).
In recent years, research on psychology, neuroscience and decision making has also
revealed that people’s emotional status determines their decision process to a great degree,
as emotion changes how people process information (Shiv et al. 2005). However, most
existing IS models or theories developed to predict and explain website adoption empha-
sised much on cognitions (e.g. perceived usefulness, perceived ease of use, etc.), while the
impact of users’emotions in human–computer interaction is traditionally neglected (Deng
and Poole 2010; Sun and Zhang 2006). This may be considered the reason why existing
theories and models are deficient in explaining online consumer behaviour (Sun and
Zhang 2006).
In recent years, IS research has therefore started to realise the important role emotion
plays in influencing users’online behaviours. Some research has conducted empirical
studies to explore how website designs influence users’emotions (Deng and Poole 2010;
Éthier et al. 2008; Mummalaneni 2005; Wang, Minor, and Wei 2011). These studies
showed that proper website visual designs could induce users’positive emotional experi-
ences. However, most of these studies mainly focused on the abstract aspects of users’
perceptions towards the whole website, such as webpage aesthetics, webpage complexity
and online store atmosphere (Deng and Poole 2010; Lavie and Tractinsky 2004). There is
a need, therefore, to study specific website design elements that facilitate a positive
emotional experience at a finer granular level (Parboteeah, Valacich, and Wells 2009),
which can lead to more operational recommendations for website designers.
Human image is one of the common visual design elements in website design.
However, research regarding the impact of human images in the context of online
shopping is particularly sparse. Recently, there have been a few studies on human
image in e-commerce websites (e.g. Cyr et al. 2009; Seo, Chae, and Lee 2012). Cyr
et al. (2009) created a construct that is image appeal and demonstrated that human image
had a positive impact on online trust through enhancing image appeal and perceived social
presence using a questionnaire, interviews and eye-tracking methodology. On the basis of
Cyr et al. (2009), Seo, Chae, and Lee (2012) examined how human brand image appeal
affected visual attention and purchase intention using eye-tracking and questionnaire
methods. However, either of the two studies did not include a direct measurement of
emotion in their questionnaires and eye-tracking methods or investigate the impact of
human image on users’emotions.
In the meantime, previous studies in the e-retail area have found that online con-
sumers’focuses and concerns during their shopping may differ according to product types
(Hsieh, Chiu, and Chiang 2005; Weathers, Sharma, and Wood 2007). However, existing
studies on human image elements on websites (Cyr et al. 2009; Seo, Chae, and Lee 2012)
neglected the impact of different product types.
Applying the Mehrabian–Russell environmental psychology model, this research
attempts to explore how human image as a visual stimulus impacts users’emotion and
subsequent attitudes towards making a purchase on a certain e-commerce website using
both questionnaire and eye-tracking methods. Specifically, this article will focus on the
following three questions: (1) whether and how human image in product pictures influ-
ences users’emotion; (2) whether users’emotion affects their following responses, such
Enterprise Information Systems 583
Downloaded by [Zhejiang University] at 02:10 13 July 2014
as attitudes towards the website; and (3) whether different product types can moderate the
effects of human image.
In the following sections of this article, a review of relevant literatures and theories
comes first in Section 2. Research model and hypotheses are developed in Section 3. In
Section 4, the research methodology is reported. Data analysis is in Section 5. The final
section discusses the results and the implications of the study as well as the limitations and
possible future research directions.
2. Theoretical background
2.1. The Mehrabian–Russell environmental psychology model
The environmental psychology model proposed by Mehrabian and Russell (1974) is used
as the theoretical framework in this study for understanding the effect of human images in
product pictures on B2C websites. This model is operationalised in a stimulus–organism–
response (S–O–R) framework, which posits that environmental stimuli affect an indivi-
dual’s cognitive and emotional reactions, which in turn affect behaviours.
Mehrabian and Russell conceptualised their model under various environments, and
this model was mostly used in the retail and service area (Machleit and Mantel 2001). In
recent years, this model has been extended to the studies of online retailing. These studies
demonstrated the effects of online shopping website designs on users’emotional
responses and approach behaviours towards the website (Mummalaneni 2005; Richard
2005; Parboteeah, Valacich, and Wells 2009). For the research purpose of this article, we
extended the Mehrabian–Russell model by drawing on the literatures on web design,
human image and emotion and proposed a research model of how human image influ-
enced users’cognitive and emotional reactions and subsequent attitudes towards online
shopping websites.
2.2. Human image as environmental stimuli
The research of webpage design, which focuses on users’emotional experiences, has
emerged in recent years (Lavie and Tractinsky 2004; Cyr, Head, and Ivanov 2006; Deng
and Poole 2010). These research aims at creating a new perspective to study webpage
design through the emotional experiences of users, challenging the original paradigm that
emphasised on webpage usability. The findings of existing research suggest that proper
webpage visual design has the potential of stimulating emotional attractions in users
(Garrett 2008).
Despite that an online shopping website consists of many stimuli that may induce the
emotions of users, this research focused on human images in product pictures. As an
important design element, human image has been studied in a few website-related
research (Hassanein and Head 2007; Cyr et al. 2009; Djamasbi, Siegel, and Tullis 2010;
Djamasbi et al. 2010; Seo, Chae, and Lee 2012). Using emotive texts and human images
can enhance the perceived social presence of users (Hassanein and Head 2007). It also has
been suggested that human image increased the visual appeal of webpages (Djamasbi
et al. 2010) and images of celebrities had positive effects on users’attitudes towards the
website (Djamasbi, Siegel, and Tullis 2010). Cyr et al. (2009) constructed the concept of
image appeal and found that human image could increase the perceived image appeal and
social presence, thus promoting users’trust towards the website. Seo, Chae, and Lee
(2012) further suggested that the high level of human brand image appeal had a positive
584 Q. Wang et al.
Downloaded by [Zhejiang University] at 02:10 13 July 2014
influence on the consumer’s visual attention and purchase intention towards a product.
These findings imply that human image can bring about positive emotional stimuli to
users. However, existing research regarding human image rarely focuses on the explora-
tion of the impact that human image exerts on users’emotions.
2.3. Theory of emotion
There are three generally accepted approaches to studying emotions in the marketing
field: categories, dimensions and cognitive appraisals (Watson and Spence 2007). The
categories approach, namely the basic emotions approach, is based on cross-cultural and
developmental research that suggests the existence of a finite set of discrete emotions-such
as joy, anger, sadness and fear, which are innate to all human beings (e.g. Izard 1992;
Plutchik 1982). The dimensions approach typically distinguishes between a dimension of
affective valence (i.e. direction) and a dimension of affective arousal (i.e. intensity)
(Bagozzi, Baumgartner, and Pieters 1998). The cognitive appraisal theory suggests that
emotional response unfolds as a multistage process consisting of (1) the antecedents of the
appraisal process, (2) the process of appraising personally relevant information and (3) the
consequences of appraisals and emotions (Johnson and Stewart 2005). Cognitive appraisal
theories offer a more complete explanation of consumers’behavioural responses to
emotions than that has emerged from either of the two preceding approaches, because
appraisal theories address the causes and consequences of emotions (Johnson and Stewart
2005). Cognitive appraisal theories suggest that emotional valence is a result of cognitive
appraisal of related information (Lazarus 1982), while information processing is an
important aspect for online shopping (Éthier et al. 2006). So this study adopted the
cognitive appraisal perspective to develop the research model. We assumed that when
exposed to a webpage with human image (stimulus), consumers will have both cognitive
and emotional reactions. A cognitive reaction relates to the appraisal of the information
about the stimulus and ultimately determines the emotional reactions to the stimulus.
In this research, we focused on positive emotions since the simulated websites did not
contain elements that may cause negative emotions. There are a number of positive
emotions, such as enjoyment (Van der Heijen 2004; Cyr, Head, and Ivanov 2006; Lin
and Bhattacherjee 2010), playfulness (Hsu and Chiu 2004; Chu and Lu 2007) and flow
(Huang 2003; Wu and Chang 2005). In this study, we chose the emotion of ‘enjoyment’
as the research target. There are reasons that justify our choice: First, as a research
construct, this positive emotion has been studied for more than 15 years in diverse
disciplines, such as business research (Bauer, Falk, and Hammerschmidt 2006), social
psychology (Davis, Bagozzi, and Warshaw 1992), retailing (Childers et al. 2001), market-
ing (Lin, Gregor, and Ewing 2008) and IS research (Van der Heijen 2004; Cyr, Head, and
Ivanov 2006). Second, it was regarded as a multidimensional emotion and has been
argued that it should be treated as a complex phenomenon (Lin, Gregor, and Ewing
2008). Compared with those basic emotions, such as joy, anger and fear, this multi-
dimensional emotion will offer more insights and opportunities for us to understand the
functions of emotions (Lin and Vasilyeva 2011).
3. Research model and hypotheses
Under the S–O–R framework, this article proposed a research model (shown in Figure 1)
to test the impact of human image in product pictures on shopping websites.
Enterprise Information Systems 585
Downloaded by [Zhejiang University] at 02:10 13 July 2014
In this research, human image in product picture was defined as the stimulus that can
induce emotional changes in users. Image appeal and perceived social presence were two
constructs used to measure users’cognitive appraisals that corresponded to stimuli, as the
perception of visual image appeal and social presence both involved cognitive processes
(Biocca, Harms, and Burgoon 2003; Karanika 2007). Image appeal refers to the extent to
which product images on the website are perceived as appropriate and aligned to user
expectations, satisfying or interesting (Cyr et al. 2009). Perceived social presence is
defined as the extent to which a medium gave users a sense of human warmth and
sociability (Hassanein and Head 2007). Divergent from the most past applications of the
S–O–R model, the direct relationship between cognitive and emotional reactions was
examined in this study. Based on the cognitive appraisal theory of emotion, we inferred
that human image will induce positive emotion in online users through image appeal and
perceived social presence. In this article, positive emotion was defined as enjoyment.
Consistent with Hassanein and Head (2007), we used attitude as the endogenous
construct rather than behavioural intention in the research model. On the one hand, this
research utilised a controlled experimental design with manipulated fictitious websites. As
such, asking participants to report their attitude may be more appropriate than their
behavioural intentions. On the other hand, attitudes can be used to predict behaviour
(Ajzen and Fishbein 1977). In online shopping context, researchers found that attitudes
did have a significantly positive effect on users’behavioural intention (Pavlou and
Fygenson 2006; Van der Heijden 2003). Therefore, this research model focused on the
indirect effect of human image on users’attitudes through its impact on users’emotions.
The research model is presented in further detail in the following sections.
3.1. Human image and image appeal
Although rarely researched, prior research has suggested that human image did affect
website browsers. Internet users between the age of 18 and 31 gave the facial area more
fixations when browsing the webpage with human image (Djamasbi, Siegel, and Tullis
2010). Cyr et al. (2009) built the construct of ‘image appeal’and demonstrated that human
image significantly increased the image appeal of pictures of electronic products. In fact,
the higher level of media richness implied a closer social and emotional connection with
H 3
H1
H2
Product Type
H5
H4
Image
Appeal
Human
Image
Perceived
Social
Presence
Enjoyment
Attitude
H 6
Stimulus Organism Response
Figure 1. Research model.
586 Q. Wang et al.
Downloaded by [Zhejiang University] at 02:10 13 July 2014
the users (Straub and Karahanna 1998). Therefore, it can be inferred that product pictures
with human images are more attractive to users than product pictures without human
images.
Cyr et al. (2009) speculated that image appeal may be context dependent and the
relationship between human image and image appeal may differ across product types.
Burke (2002) proposed an entertainment–utilitarian product classification in which enter-
tainment products were defined as products that consumers pay more attention to their
style and want to have fun and entertaining shopping experiences while utilitarian
products referred to products whose detailed product information such as functional
indicator is more important to consumers and consumers buy them to use certain func-
tions. It can be deduced that for an entertainment product, the presence of human image
would facilitate users to have better product experience, as users pay much attention to
product style, while for a utilitarian product, the embedment of human image in its
product picture may not cause a significant effect, as users may be more concerned
about a product’s functional features. Therefore, we believe that this product classification
can moderate the relationship between human image and image appeal. Thus, the follow-
ing hypotheses can be derived.
Hypothesis 1: Human image in product picture can improve the appeal of the
product picture significantly.
Hypothesis 2: Product type will moderate the effect of human image in product
picture on image appeal, namely the relationship between human image and product
image appeal is stronger for websites selling entertainment products than for websites
selling utilitarian products.
3.2. Human image and perceived social presence
Integrating human image and proper text on a website can induce social presence in users,
conveying a feeling of others’presence (Gefen and Straub 2004). Research results showed
that when human image and socially rich text were used, the perceived social presence
was significantly higher than conditions using functional text and a plain product picture
(Hassanein and Head 2006, 2007). The research result of Cyr et al. (2009) also showed a
significant positive relationship between human image and users’perceived social pre-
sence. Based on previous studies, it can be expected that websites using product picture
combined with human image will have a higher level of human contact and social
presence than websites using product picture without human image. Therefore, the
following hypothesis can be derived.
Hypothesis 3: Product picture with human image can improve the perceived social
presence of websites significantly.
3.3. Image appeal, social presence and enjoyment
According to the cognitive appraisal theory of emotion, cognition determines emotion,
which ultimately impacts behaviour (Holbrook and Batra 1987). More specifically, when
exposed to a stimulus, an individual processes information about the stimulus to develop
appraisals that ultimately determine the emotional reactions to the stimulus (Berkowitz
Enterprise Information Systems 587
Downloaded by [Zhejiang University] at 02:10 13 July 2014
1993). The cognitive reactions to the environment can thus have an enhancing or a
deterring effect on the emotional reactions experienced and would be viewed as an
antecedent to emotion (Parboteeah, Valacich, and Wells 2009). The relationship between
cognition and emotion has been empirically examined in IS research (Van der Heijen
2004; Cyr, Head, and Ivanov 2006; Parboteeah, Valacich, and Wells 2009). For example,
the results of Parboteeah, Valacich, and Wells (2009) found that interaction with a website
led to both cognitive reactions (perceived usefulness) and affective reactions (enjoyment),
and perceived usefulness had a positive effect on enjoyment.
In this study, image appeal and perceived social presence were two constructs used to
measure users’cognitive appraisals that corresponded to stimuli, as the perceptions of
visual image appeal and social presence both involved cognitive processes (Biocca,
Harms, and Burgoon 2003; Karanika 2007). Prior to bringing about enjoyment, human
image was first automatically recognised as processing attractive or possessing social
presence. Existing research has implied the positive relationship between human image
and enjoyment from the perspective of social presence (Swan and Shih 2005) and visual
appeal (Lindgaard et al. 2006). Image appeal affects the aesthetics of a website signifi-
cantly. Research showed the aesthetics of the website influenced users’emotion (Lavie
and Tractinsky 2004; Cyr, Head, and Ivanov 2006); specifically, design aesthetics was
positively related to enjoyment (Cyr, Head, and Ivanov 2006). Websites perceived to have
a high level of social presence also improve perceived enjoyment because the impacts of
medium on the emotions and behaviours of users also increase when the social presence
in the medium increases (Argo, Dahl, and Manchanda 2005). Researchers have demon-
strated that perceived social presence can affect users’perceived enjoyment positively
(Qiu and Benbasat 2009; Hassanein and Head 2007).
Therefore, we believe that the relationship between human image and enjoyment is
indirect and image appeal and perceived social presence play the mediator role. Hence, the
following hypotheses can be derived.
Hypothesis 4: Higher image appeal will result in greater enjoyment in online shop-
ping websites.
Hypothesis 5: Higher perceived social presence will result in greater enjoyment in
online shopping websites.
3.4. Enjoyment and attitude towards online shopping websites
According to the S–O–R paradigm, the emotional reactions to the environment will
determine an individual’s response (Mehrabian and Russell 1974), which, in this study,
is the attitude towards online shopping websites. Using the S–O–R paradigm as a
foundation, several studies on traditional environment have found empirical evidence
supporting the positive relationship between enjoyment or pleasure and several approach
behaviours, such as time spent in the store, money spent in the store, purchase intentions
and impulse purchases (Baker, Levy, and Grewal 1992; Sherman, Mathur, and Smith
1997; Turley and Milliman 2000; Yalch and Spangenberg 2000). Studies in IS have also
shown that emotions can play the same role in online environment. These studies found
that enjoyment was an antecedent to user attitudes, such as M-Loyalty (Cyr, Head, and
Ivanov 2006), E-Loyalty (Cyr et al. 2007), E-Trust (Hwang and Kim 2007) and user
attitudes towards hedonic systems (Lin and Bhattacherjee 2010).Therefore, the following
hypothesis can be derived.
588 Q. Wang et al.
Downloaded by [Zhejiang University] at 02:10 13 July 2014
Hypothesis 6: The stronger the enjoyment users feel, the more positive their attitudes
towards the online shopping websites are.
4. Research methodology
The methods of emotion measurement can be roughly divided into neuro-physiological
signal processing methods, observer methods and self-report methods (Lopatovska and
Arapakis 2011). Each of these methods has its own advantages and disadvantages. It is
suggested that researchers should consider using these methods in combination to
improve the reliability of findings and represent users’affective states more compre-
hensively (Lopatovska and Arapakis 2011). Eye tracking is a kind of neuro-physiolo-
gical method that can record the eye-movement metrics to objectively reflect
participants’attentions and emotions during the experiment. As an eye metric, pupil
size variance can reflect the emotion change of people (Hess and Polt. 1960; Partala and
Surakka 2003; Bradley et al. 2008). So this study combined questionnaire (self-report
method) and eye-tracking (neuro-physiological method) to measure users’emotions and
investigate our hypotheses.
A 2 × 2 mixed laboratory experiment was designed to test the research model. The
between-subjects factor is product picture design: product picture with human image and
product picture without human image. The within-subjects factor is product type: enter-
tainment product and utilitarian product. In line with Hassanein and Head (2006), this
research used apparel to represent entertainment products and headphones to represent
utilitarian products. According to the data from CNNIC (http://www.cnnic.cn/hlwfzyj/
hlwxzbg/), apparel and digital products are currently the two types of products that have
the largest domestic online transactions and people are therefore more familiar with them.
Besides, these two types of products are also quite representative in the product classifica-
tion proposed by Burke (2002).
4.1. Experimental apparatus and materials
4.1.1. Experimental webpage
The four websites used in the main experiment were constructed using Microsoft Visual
Studio 2008. Two of them displayed the apparel products and the other two websites
displayed the headphones products. Furthermore, on the two websites with the same type
of products, product pictures on one website had human elements while the other website
did not. On each website, three webpages displayed three products of the same type with
the same level of human elements. Each product was displayed with a concise text
description and two pictures (one from the front and the other from the back or side).
All the design elements on the four websites were the same except specific product names,
pictures and descriptions. As this research used two different models to display the
apparel products and the headphones, a test about the attractive level of the two models
was conducted. Forty-four people were shown pictures of the two models and asked to
rate their attractiveness on a 7-point Likert scale. A paired sample t-test showed that there
was no significant difference between the two models’attractive levels (p> 0.05,
t= 1.296). The websites were designed to be concise and neutral, thus minimising the
possibility that users’emotions would be affected by design elements irrelevant to the
research topic. The colour and layout were designed based on the current leading B2C
online shopping websites while a fictitious website name was used (called EasyGo.com)
Enterprise Information Systems 589
Downloaded by [Zhejiang University] at 02:10 13 July 2014
to avoid the latent brand and reputation effects. In the experiment, each participant
browsed two websites that had the same human image level but different product type.
4.1.2. Eye-tracker
The eye-tracking device used in this research was the iView X Hi-Speed eye-tracker
produced by the German company SMI. Its sampling rate is 500 HZ. The infrared camera
is used in the Hi-Speed eye-tracking equipment to capture the video image of participants’
left eye. Signal export from the camera is first coded using MPEG coding rule and then
entered in the computer for the collection and analysis of image data, so that the moving
distance, speed, pupil diameter and fixation position of the eye can be calculated in real
time. The experimental websites were presented in a 19-inch monitor with a resolution of
1024*768 pixels. During the experiment, participants were asked to put their chins on the
chin rest and to adjust their body to a comfortable sitting position and keep their head
static. The distance between monitor and chin rest was approximately 60 cm. The eye-
tracking data of websites browsing in the experiment was recorded automatically by the
eye-tracker and analysed using the software BeGaze 2.5 matched to the eye-tracker.
4.2. Experimental procedure
We conducted a pilot experiment before the formal experiment. Four participants took part
in the pilot experiment. Thus, we can standardise the experiment processes and guarantee
the appropriateness of the experiment settings. In the formal experiment, due to the nature
of the experiment (i.e. using an eye-tracking device), the lab was reserved for one
participant at a time. Participants first read and signed an informed consent form for the
eye-tracking experiment, and then read the experiment instruction in the waiting room.
We described the research goal as ‘user experience in online shopping websites’in the
instruction. The experimental task was described as follows: ‘EasesyGo.com is an online
shopping website. It has two sub-websites that sell female apparels and women head-
phones respectively. You need to visit these two sub-websites later and select a T-shirt and
a pair of headphone for a female friend of yours. There are three T-shirts and three pairs of
headphones in the two sub-websites respectively for you to choose. There is no time
limitation in terms of website browsing and purchase’. The experiment started only after
participants verbalised full understandings about the experimental procedure and task.
The formal experiment was conducted in a soundproof and dimly lit laboratory.
Participants were calibrated to the eye-tracker before browsing the websites. Calibration
took 3 minutes on average. Participants who wore glasses normally took a longer time to
be calibrated, as the reflection of their glasses needed to be offset by carefully adjusting
the eye-tracker. All participants were calibrated successfully. After the calibration, parti-
cipants began to browse the websites. Each participant browsed two websites in total. For
each participant, the two websites displayed two types of product but with the same level
of human image elements. Participants browsed two websites in a random order to avoid
the latent order effect. Each website had three same types of products and participants
could take their time for the task. The eye-tracker recorded the information of participants’
eye movements during the entire browsing process.
The task of browsing the first website terminated after the participants purchased the
first product. Then the participants filled in a questionnaire in the waiting room and
relaxed their eyes for five minutes. After the break, participants were calibrated again and
then started the task of browsing the second website. Likewise, they filled in the same
590 Q. Wang et al.
Downloaded by [Zhejiang University] at 02:10 13 July 2014
questionnaire when they finished this second task of the experiment. The whole experi-
ment lasted for approximately 20 minutes. Participants filled in their personal information
before they left, including demographic characteristics, Internet surfing frequency, online
shopping experiences, etc. Each participant was paid 25 Yuan (RMB) for participation.
4.3. Questionnaire design
A questionnaire was administered after each participant finished browsing a website. The
questionnaire employed the 7-point Likert scale from 1 (completely disagree) to 7
(completely agree). We analysed the questionnaire data with SPSS 16.0 software and
SmartPLS 2.0 for structural equation modelling. All constructs in the questionnaire were
built using or adapting previously developed and validated scales. These measurement
items and their sources are presented in Table 1.
4.4. Participants
Thirty-nine participants took part in the experiment. The participants were recruited from
an online discussion forum affiliated with a university in southern China. We believe that
students are the appropriate participants for this experiment as this group represents the
majority of Chinese Internet users at present (CNNIC 2012). All participants had normal
or corrected-to-normal vision. Students whose left eye visions were above 350 degrees of
myopia and with astigmatism were not recruited. They were randomly assigned to
different experiment conditions. All of them had online shopping experiences.
Participants consisted of 19 females and 20 males, most between the age of 22–24
(82.1%). 76.9% of the participants surfed the Internet for 1–5 hours per day on average.
Table 1. Measurement items of constructs.
Constructs Measurement Items Source
Image Appeal (IA) The product images used in the website are
appropriate
Cyr et al. (2009)
The product images used in the website are satisfying
The product images used in the website are exciting
The product images used in the website are interesting
The product images used in the website make the
products look appealing
Perceived Social
Presence (SP)
There is a sense of human contact on this website Gefen and Straub
(2003)There is a sense of sociability on this website
There is a sense of human warmth on this website
Enjoyment (E) I found my visit to this website interesting Hassanein and
Head (2006)I found my visit to this website entertaining
I found my visit to this website enjoyable
I found my visit to this website pleasant
Attitude (AT) I would have positive feelings towards buying a
product from this website
Van der Heijden
(2003)
The thought of buying a product from this website is
appealing to me
It would be a good idea to buy a product from this
website
Enterprise Information Systems 591
Downloaded by [Zhejiang University] at 02:10 13 July 2014
76.5% of the participants had at least 2–6 years of online shopping experience. Nineteen
of the participants were assigned to the control group (no human image) and the other 20
were assigned to the experimental group (with human image). An independent samples
t-test showed that there were no significant differences between participants in the control
group and the experimental group in terms of average Internet surfing time per day
(p= .247) and online shopping experience (p= .395). Likewise, no significant differences
were found for demographic characteristics such as gender (p= .276), age (p= .990) and
education (p= .661).
Moreover, as the human image used in the experimental websites were female figures,
we were concerned about the image appeal might be influenced by the gender of
participants. However, the independent samples t-test of the experimental group divided
by gender showed that there was no significant gender difference in image appeal of the
product picture with human image (p= .607).
5. Data analysis and results
The principal data analysis method employed in the study was Structural Equation
Modelling (SEM), and the results of SEM were supplemented by the eye-tracking data
analysis. The SEM method can examine the measurement relationship between factors as
well as the structural relationship between factors, and provide a more comprehensive
analysis of the causal relationships (Lin 2008). The variance-based PLS (Partial Least
Squares) method was chosen over covariance-based methods such as LISREL in this
study for the following reasons: (1) PLS is relatively robust to deviations from a multi-
variate distribution; and (2) PLS is suitable for analysing small samples (Chin, Marcolin,
and Newsted 2003). According to the advice from Gefen, Straub, and Boudreau (2000),
our sample size of 78 meets the requirement of sample size for PLS estimation
procedures.
5.1. Analysis of reliability and validity
Although all of the constructs in this study were measured using previously developed
and validated scales, quality assessment of the final dataset can provide further verifica-
tions. Composite Reliability (CR), Average Variance Extracted (AVE) and Cronbach’s
alpha were used as indicators to test construct reliability in PLS. The results are presented
in Table 2. The Cronbach’s alpha for each construct was higher than the recommended
level of 0.70 (Rivard and Huff 1988). The composite reliability (CR) of all the latent
variables was higher than the recommended level of 0.60 (Bagozzi and Yi 1988). The
average variance extracted (AVE) value for each construct was higher than the recom-
mended level of 0.50 (Bagozzi and Yi 1988). All of these results indicated good
reliability.
As shown in Table 2, all of the measurement items loaded on their respective factors
with strong statistical significance (p< 0.01), indicating good convergent validity. In
addition, the variance-extracted test was used to establish discriminant and convergent
validity. Validity is demonstrated if the square root of the AVE of each construct is higher
than the correlations between it and other constructs (Fornell and Larcker 1981). The
results (shown in Table 3) indicated good convergent and discriminant validity.
592 Q. Wang et al.
Downloaded by [Zhejiang University] at 02:10 13 July 2014
5.2. Hypothesis test
Figure 2 presents the results of PLS analysis of the research model. Bootstrapping
procedures were performed to determine the path coefficients and statistical significance
of each hypothesised path, with the bootstrap resamplings set at 500. Human image and
product type were entered in the model as nominal variables; this is reasonable in PLS
modelling (Henseler and Fassott 2010) and has been adopted in previous research (Cyr
et al. 2009). In terms of human image, ‘0’represented the condition without human image
whereas ‘1’represented that with human image. Likewise, in terms of product type, ‘0’
represented apparels whereas ‘1’represented headphones. Results showed that all the path
coefficients were statistically significant (p< 0.05).
The PLS product-indicator approach was used to determine the moderating effect in
the research model. The path coefficient of the moderating effect of product type on the
relationship between human image and image appeal was −0.309. This showed that when
the statistical value of product type increased (from apparels to headphones), the impact of
Table 2. Factor loading of indicator and construct, Tvalue, composite reliability, AVE and
Cronbach’s alpha.
Construct and indicator
Factor
Loading T
Composite
reliability AVE
Cronbach’s
alpha
Image Appeal (IA) 0.90 0.65 0.86
IA1 0.63 7.84
IA2 0.81 18.86
IA3 0.80 19.22
IA4 0.88 28.46
IA5 0.89 40.84
Perceived Social Presence (SP) 0.91 0.78 0.86
SP1 0.88 32.66
SP2 0.87 26.26
SP3 0.90 39.73
Enjoyment (E) 0.88 0.65 0.82
E1 0.83 28.02
E2 0.73 13.26
E3 0.78 12.70
E4 0.88 46.49
Attitude (AT) 0.94 0.83 0.90
AT1 0.88 36.18
AT2 0.93 56.70
AT3 0.92 67.62
Table 3. Correlation matrix and the square root of AVE.
Image Appeal Perceived Social Presence Enjoyment Attitude
Image Appeal 0.81
Perceived Social Presence 0.36 0.88
Enjoyment 0.70 0.53 0.80
Attitude 0.72 0.51 0.70 0.91
Note: The diagonal elements are the square root of AVE.
Enterprise Information Systems 593
Downloaded by [Zhejiang University] at 02:10 13 July 2014
human image on image appeal declined. The independent samples t-test on image appeal
regarding apparels and headphones further illustrated the moderating effect of product
type. As shown in Table 4, integrating human image in apparel pictures increased the
image appeal significantly (p< 0.01) while integrating human image in headphones
pictures had no significant impact on image appeal (p= 0.682).
Hence, all the hypotheses (H1–H6) were supported. The summary of the path
coefficients and the corresponding t-statistics for each hypothesised path in the model is
presented in Table 5. Moreover, 48.9% of the variance in the attitude towards websites
was explained by the variables in the model. The R
2
of the other endogenous constructs
such as image appeal, perceived social presence and enjoyment were 21.4%, 12.2% and
57.7%, respectively. This shows a good predictive capability of the model.
Table 4. Independent samples t-test on image appeal of apparels and headphones.
Variable Group
Subject
number mean
Standard
deviation tSig.
Image appeal (apparel) Without human
image
19 4.516 1.106 −3.882 .000
With human image 20 5.720 .817
Image appeal
(headphone)
Without human
image
19 4.684 1.018 .412 .682
With human image 20 4.550 1.013
Table 5. Results of hypotheses testing.
Hypothesis Causal path Path coefficient t-statistic Supported
H1 human image –> image appeal 0.248 2.925** yes
H2 product type * human image –> image appeal −0.309 3.508** yes
H3 human image –> perceived social presence 0.349 4.290** yes
H4 image appeal –> enjoyment 0.582 11.149** yes
H5 Perceived social presence –> enjoyment 0.324 5.192** yes
H6 enjoyment –> attitude 0.699 14.362** yes
Note: **p< 0.01.
–0.239* –0.309**
0.699**
0.324**
0.582**
0.349**
0.248**
Image Appeal
R2 = 0.214
Human Image Enjoyment
R2 = 0.577
Attitude
R2 = 0.489
Product Type Product Type
*Human Image
Perceived Social
Presence
R2 = 0.122
*: p-value < 0.05
**: p-value < 0.01
Figure 2. PLS structure model.
594 Q. Wang et al.
Downloaded by [Zhejiang University] at 02:10 13 July 2014
5.3. Moderating effect size
The moderating effect of product type has been demonstrated through the statistical
significance of the path coefficient (p< 0.01). Furthermore, the moderating effect size
can be deduced by comparing the percentages of variance explained by the model in
relation to the moderated variable without and with the presence of the moderating variable
(Chin, Marcolin, and Newsted 2003). In our model, the former was 0.214 while the latter
was 0.062. The moderating effect f
2
thus can be calculated by the following formula:
f2¼
R2
model with moderating variable R2
model without moderating variable
1R2
model with moderating variable
The overall effect sizes f
2
for the interaction of 0.02, 0.15 and 0.35 have been suggested
to be small, moderate and large effects, respectively (Chin, Marcolin, and Newsted 2003).
The moderating effect size of product type was f
2
= 0.193, which suggested a moderate
effect.
5.4. Mediating effects test
Image appeal and perceived social presence can be viewed as cognitive variables of
organism part in the S–O–R framework in this research. To test the mediating effects of
the cognitive process, we first established a new model in which the mediating variables
(image appeal and perceived social presence) were removed. Then the two mediating
variables were added back to the above model and we compared the statistical signifi-
cance of the path coefficients linked from human image to enjoyment in the two models.
This method to test for mediation was proposed by Baron and Kenny (1986).
In the new model without the mediating variables, the path coefficient from human
image to enjoyment was 0.244, which was significant (p<0.01,t-value = 2.591). When the
mediating variables were added back to the model, the path coefficient from human image
to enjoyment changed from 0.244 to −0.017, which was not significant (t-value = 0.236).
Besides, the variance of enjoyment that can be explained was 57.7% in the original model
while in the model without mediating variables only 5.9% of the variance can be explained.
Based on the above findings, we believe that the mediating effects of image appeal and
perceived social presence as the cognitive process were established.
5.5. Analysis of eye-tracking data
5.5.1. Pupil size
Change in pupil size can reflect the emotional change of users. Pupil responds to emotion
and dilates when people see things they like. Eye-tracker was employed in this study to
record the pupil size of users so that it became possible to compare pupil sizes under
different experimental conditions. This research assumed that human image in product
picture can engender pleasant emotions from users. The product picture areas under two
experimental conditions (with/without human image) were defined as the areas of interest
(AOI). We obtained all the eye-relevant information in the AOI. An independent samples
t-test with the data of pupil sizes was conducted to test whether participants’pupil sizes
will be significantly larger in the presence of human image than those in the absence of
human image. The data of two product types was analysed, respectively. Table 6 shows
the statistical results.
Enterprise Information Systems 595
Downloaded by [Zhejiang University] at 02:10 13 July 2014
Results showed that regardless of product types, the presence of human image can
increase the pupil size in both crosswise width (i.e. pupil size X) and lengthways width
(i.e. pupil size Y) significantly (p< 0.01) compared to the conditions without human
image. Thus, we believe that the analysis results of pupil size supported the research
hypothesis, namely human image is a positive emotional stimulus for users.
5.5.2. Fixation count and fixation time
Fixation information can be used to measure the attention that individuals have paid to
stimuli (Vertegaal and Ding 2002) and a higher level of attention usually indicates a
higher level of attraction. In this research, we compared the fixation indicators in product
picture between the control group and the experimental group to explore whether the
image appeal would be higher in the presence of human image. We analysed the eye-
tracking data of the product picture AOI. The related fixation indicators included fixation
count, accumulated fixation time and fixation time percentage in total browsing time.
Tables 7 and 8 show the results of the independent samples t-test of the fixation indicators
of the two product types.
Table 6. Independent samples t-test of pupil size.
Variable (Pupil size) group Mean (mm) S.D. tSig.
X (apparel) Without human image 39.139 5.138 −22.538 .000
With human image 45.515 8.040
Y (apparel) Without human image 39.032 5.453 −21.491 .000
With human image 45.047 7.586
X (headphones) Without human image 38.744 5.937 −16.375 .000
With human image 44.036 7.996
Y (headphones) Without human image 39.141 6.290 −14.997 .000
With human image 43.823 7.259
Table 7. Independent samples t-test of fixation indicators (apparel).
Variables group mean S.D. tSig.
Fixation count Without human image 9.840 5.692 −3.793 .000
With human image 13.925 8.707
Fixation time (ms) Without human image 2.637E3 1747.66 −2.961 .004
With human image 3.541E3 2378.79
Fixation percentage Without human image 30.513 13.995 −3.257 .001
With human image 37.766 16.375
Table 8. Independent samples t-test of fixation indicators (headphones).
variable group mean Std deviation tSig.
Fixation count Without human image 12.788 7.691 −.866 .387
With human image 13.782 8.513
Fixation time (ms) Without human image 3.629E3 2348.779 −.797 .426
With human image 3.897E3 2409.159
Fixation percentage Without human image 37.508 18.467 −.175 .861
With human image 37.059 17.789
596 Q. Wang et al.
Downloaded by [Zhejiang University] at 02:10 13 July 2014
The results showed that for apparel product, the fixation count, fixation time and
fixation time percentage were significantly higher when human image was integrated into
product picture (p< 0.01). So we believe that the analysis results of fixation data
demonstrated that for apparel product (entertainment product), integrating human image
in product picture can draw more attention from users (i.e. increasing the image appeal of
product picture), while for headphones (utilitarian product), integrating human image in
product picture would not increase the image appeal of product picture significantly. The
results of eye-tracking data analysis supported the hypothesis about the moderating effect
of product type on the relationship between human image and image appeal in the SEM.
5.5.3. Heat map
Heat map allowed us a more intuitive understanding of fixation results. It has been used as
an analysis tool in most papers containing the analysis of eye-tracking data (Djamasbi,
Siegel, and Tullis 2010; Djamasbi et al. 2010; Seo, Chae, and Lee 2012) and can be
interpreted as the degree of visual attention. The cumulative fixation data was translated
into a heat map and the colours on it represent the degree of users’fixations. Red indicates
the highest level of fixation, followed by yellow and green. Areas with no colour mean
receiving no fixation. The red solid diamonds in the heat maps indicate the users’mouse
clicks and are thus irrelevant to the fixation analysis.
For apparel product, the examples of fixation heat maps are shown in Figures 3 and 4.
As the figures show, participants paid more attention to product picture integrated with
human image. While for headphones, in conditions either with or without human image,
participants paid much more attention to the functional information of headphones than
the product picture (shown in Figures 5 and 6).
Figure 3. Sample of heat map for apparel product (with human image).
Enterprise Information Systems 597
Downloaded by [Zhejiang University] at 02:10 13 July 2014
6. Discussion
6.1. Main conclusions
This study used the data obtained in the eye-tracking experiment and the following
questionnaires to test the theoretical model and hypotheses. The results showed that all
Figure 5. Sample of heat map for headphones product (with human image).
Figure 4. Sample of heat map of apparel product (no human image).
598 Q. Wang et al.
Downloaded by [Zhejiang University] at 02:10 13 July 2014
hypotheses were supported. Product picture integrated with human image can result in
higher levels of image appeal and perceived social presence, and consequently stimulate
more pleasant feelings in users. This research also found that product type moderated the
relationship between human image and image appeal. Human image can significantly
increase image appeal when integrated in entertainment product while its influence was
not significant when applied to utilitarian product. This result confirmed the assumption of
Cyr et al. (2009) that image appeal may be context dependent.
The analysis results of eye-tracking data further supported the analysis results of our
theoretical model, and offered possible explanations. The pupil sizes of participants in the
experimental group during the process of viewing apparels’or headphones’product
pictures with human image were significantly larger than those of participants in the
control group. This result showed that participants in the experimental group experienced
more pleasant emotions when browsing product pictures (with human image).
It’s worth noting that the meaning of pupil size remains controversial. Although many
research show that pupil size is related to emotional status (Partala and Surakka 2003;
Bradley et al. 2008; Urry et al. 2009), pupil size can be also affected by many other
factors, such as cognitive workload (Urry et al. 2009), ambient light (Cegarra and
Chevalier. 2008) and the absence of expected stimuli (Karatekin 2007). Since the tasks
in each experiment condition (with or without human image) were simple (choose one
product out of three) and of similar difficulty levels, and the ambient light in the
laboratory has been kept constant in all the experiments, the authors believe the pupil
size variance in this research was explained by the emotional changes of the participants.
Existing literatures have shown that there is a U-shape relationship between the valence of
emotion and the pupil size (Partala and Surakka 2003). That is, pupil dilates in both
positive and negative emotions. Nowadays, it is widely accepted that emotion comprises
two parts –valence and arousal, and the dilation of pupil size only reflects the arousal
aspect of emotion (Bradley et al. 2008; Urry et al. 2009). In this research, the simulated
Figure 6. Sample of heat map for headphones product (no human image).
Enterprise Information Systems 599
Downloaded by [Zhejiang University] at 02:10 13 July 2014
websites did not contain elements that may cause negative emotions, and at the same time
the participants reported a significantly higher level of enjoyment when human images
were presented. The authors believe that it is reasonable to explain the pupil size variance
in the research as positive emotions induced by the presence of human image.
We also analysed the fixation indicators recorded by the eye-tracker. The results
showed that human image could significantly improve fixation indicators (i.e. fixation
count, fixation time and fixation time percentage) in pictures of apparel products, whereas
no significant differences were found in pictures of headphone products. We believe that a
possible explanation is that users paid much attention to the functional information of
utilitarian products. The heat maps supported this explanation. Through the visual dis-
plays of heat maps, we can see that for the headphone, its product picture was fixated
upon at a relatively low level regardless of whether human image was embedded or not,
while its functional information was fixated upon for a much longer time. This also
supported the moderating effect of product type on the relationship between human image
and image appeal, which was suggested in the theoretical model. As the eye movement
indicators and heat maps revealed, users’attention for utilitarian product was not on its
product picture. Even when its product picture was integrated with human image, users
can hardly notice its latent picture appeal.
6.2. Theoretical contributions and practical implications
Recently, the role of human image as a visual element in webpage has been explored in a
few research (Cyr et al. 2009; Hassanein and Head 2006, 2007; Seo, Chae, and Lee 2012).
However, few studies focused on the direct emotional impact that human image exerts on
users, and few measured users’emotions directly. This study investigated in detail how
human image in product picture stimulated users’emotions and the subsequent positive
responses of users. Emotion was measured through both questionnaire and eye-tracking
method. Compared with existing literatures (e.g. Cyr et al.), this research measured pupil
size to indicate participants’emotional changes, thus contributing to a more comprehen-
sive understanding of how human image as a webpage design element affects the
emotions of users. As the result showed, human image engendered positive emotion
experiences (dilated pupil size and more positive self-reported enjoyment) from users
through increasing the image appeal and social presence that users’perceived. This
provides us a way to measure users’emotions directly and to explore the potential
mechanism of how emotion is induced and takes effect. This is one of the important
theoretical contributions of this article.
Another important theoretical contribution of this research is that it demonstrated the
moderating role of product type on the relationship between human image and image
appeal, which has not been researched thoroughly before. Researchers classified e-com-
merce products according to different standards, such as entertainment products and
utilitarian products (Burke 2002), and search products and experience products (Nelson
1974). Online consumers’focuses and concerns during online shopping may differ based
on product types, and researchers believe that it is necessary to conduct comparative study
across product types (Cyr et al. 2009; Éthier et al. 2008; Hassanein and Head 2007). Some
research have attempted to study the moderating role of product types (Hassanein and
Head 2006), but its impact on the relationship between human image and image appeal
was not testified. In the study about human elements in website design, Cyr et al. (2009)
pointed out that image appeal was context dependent (e.g. product type) and this should
600 Q. Wang et al.
Downloaded by [Zhejiang University] at 02:10 13 July 2014
be studied in future. This research supported this speculation and demonstrated that
product type played a crucial role in website design.
As to research methodology, most research about website design has used ques-
tionnaires as the most popular data collection method. This study employed a multi-
method approach, combining questionnaire with eye-tracking method. Compared with
the research of Cyr et al. (2009) and Seo, Chae, and Lee (2012), this study used the eye-
tracking method to measure not only visual attention but also emotion. Through the
analysis of the eye-tracking data, we can represent the emotional status of users more
comprehensively. The analysis results of eye-tracking data provided both supplement
and validation for the test results of the theoretical model. This research is also an
instructive and exploratory attempt for a multi-method approach.
The findings of this study have important practical implications for B2C website
designers. Website designers should consider the embedment of human elements in
product pictures according to online product types. For websites selling entertainment
products such as apparels, adding human elements in the product picture is beneficial and
designers can use human elements to increase the attractiveness of product image and
perceived social presence of users and thus induce pleasant feelings eventually. This
positive influence on emotion can improve the users’attitude towards the websites, thus
potentially promoting its sales performance. For websites selling utilitarian products,
however, the practical advice is different. For the utilitarian products, users focus more
on the functional indicators that can help them compare among various products.
Therefore, instead of integrating human image in the product picture, website designers
should pay more attention to the descriptions of functional information for utilitarian
products. When website aesthetics is considered solely, adding human image in the
product picture of utilitarian products cannot have significant impacts on users, as users
do not pay much attention to the pictures of utilitarian products.
6.3. Limitations and future research directions
There are some limitations to this study and further research is necessary in the future.
First, although the sample size used in this research met the recommended thresholds for
PLS analysis, the sample size in this study was relatively small. Second, the model’s body
parts exposed to participants in different experimental conditions were different. While
only the face of the model can be seen in the headphone pictures, the upper body of the
model can be seen in the apparel pictures, which may have an influence on the visual
attention of participants. However, we believe that our experimental website setting was
reasonable as these product pictures of apparels and headphones resembled those in actual
online environments. Finally, we only used one product classification scheme and only
chose one representative product from each type of classification. Other product classifi-
cation schemes and more products should be investigated in the future to explore product
attributes that can induce users’positive emotions.
Funding
This research was supported by the grant from National Natural Science Foundation of China [grant
number 71272167], [grant number 71302122]; Ministry of Education of the People’s Republic of
China [grant number 11YJA630130].
Enterprise Information Systems 601
Downloaded by [Zhejiang University] at 02:10 13 July 2014
References
Ajzen, I., and M. Fishbein. 1977. “Attitude-Behavior Relations: A Theoretical Analysis and Review
of Empirical Research.”Psychological Bulletin 84 (5): 888–918. doi:10.1037/0033-
2909.84.5.888.
Argo, J. J., D. W. Dahl, and R. V. Manchanda. 2005. “The Influence of a Mere Social Presence in a
Retail Context.”Journal of Consumer Research 32 (2): 207–212. doi:10.1086/432230.
Bagozzi, R. P., H. Baumgartner, and R. Pieters. 1998. “Goal-Directed Emotions.”Cognition &
Emotion 12 (1): 1–26. doi:10.1080/026999398379754.
Bagozzi, R. P., and Y. Yi. 1988. “On the Evaluation of Structural Equation Models.”Journal of the
Academy of Marketing Science 16 (1): 74–94. doi:10.1007/BF02723327.
Baker, J., M. Levy, and D. Grewal. 1992. “An Experimental Approach to Making Retail Store
Environmental Decisions.”Journal of Retailing 68: 445–458.
Baron, R. M., and D. A. Kenny. 1986. “The Moderator–Mediator Variable Distinction in Social
Psychological Research: Conceptual, Strategic, and Statistical Considerations.”Journal of
Personality and Social Psychology 51 (6): 1173–1182. doi:10.1037/0022-3514.51.6.1173.
Bauer, H. H., T. Falk, and M. Hammerschmidt. 2006. “Etransqual: A Transaction Processbased
Approach for Capturing Service Quality in Online Shopping.”Journal of Business Research 59
(7): 866–875. doi:10.1016/j.jbusres.2006.01.021.
Berkowitz, L. 1993. “Towards a General Theory of Anger and Emotional Aggression: Implications
of the Cognitive-Neoassociationsitic Perspective for the Analysis of Anger and Other
Emotions.”In Advances in Social Cognition, Vol. 6, edited by R. S. Wyer and T. K. Srull, 1–
46. Hillsdale, NJ: Erlbaum.
Biocca, F., C. Harms, and J. K. Burgoon. 2003. “Toward a More Robust Theory and Measure of
Social Presence: Review and Suggested Criteria.”Presence: Teleoperators and Virtual
Environments 12 (5): 456–480. doi:10.1162/105474603322761270.
Bradley, M. M., L. Miccoli, M. A. Escrig, and P. J. Lang. 2008. “The Pupil as a Measure of
Emotional Arousal and Autonomic Activation.”Psychophysiology 45 (4): 602–607.
doi:10.1111/j.1469-8986.2008.00654.x.
Bucy, E. 2000. “Social Access to the Internet.”The Harvard International Journal of Press/Politics
5 (1): 50–61. doi:10.1162/108118000568967.
Burke, R. R. 2002. “Technology and the Customer Interface: What Consumers Want in the Physical
and Virtual Store.”Journal of the Academy of Marketing Science 30 (4): 411–432. doi:10.1177/
009207002236914.
Cegarra, J., and A. Chevalier. 2008. “The Use of Tholos Software for Combining Measures of
Mental Workload: Toward Theoretical and Methodological Improvements.”Behavior Research
Methods 40 (4): 988–1000. doi:10.3758/BRM.40.4.988.
Childers, T. L., C. L. Carr, J. Peckc, and S. Carson. 2001. “Hedonic and Utilitarian Motivations for
Online Retail Shopping Behavior.”Journal of Retailing 77 (4): 511–535. doi:10.1016/S0022-
4359(01)00056-2.
Chin, W. W., B. L. Marcolin, and P. R. Newsted 2003. “A Partial Least Squares Latent Variable
Modeling Approach for Measuring Interaction Effects: Results from a Monte Carlo Simulation
Study and an Electronic-Mail Emotion/Adoption Study.”Information Systems Research, 14(11),
189–217.
Chu, C. W., and H. P. Lu. 2007. “Factors Influencing Online Music Purchase Intention in Taiwan:
An Empirical Study Based on the Value-Intention Framework.”Internet Research 17 (2): 139–
155. doi:10.1108/10662240710737004.
CNNIC. 2012. Statistical Survey Report on The Internet Development in China. http://www.cnnic.
net.cn/hlwfzyj/hlwxzbg/201201/P020120709345264469680.pdf, 2012-01-16/2014-06-03.
Cyr, D., K. Hassanein, M. Head, and A. Ivanov. 2007. “The Role of Social Presence in Establishing
Loyalty in E-Service Environments.”Interacting with Computers 19 (1): 43–56. doi:10.1016/j.
intcom.2006.07.010.
Cyr, D., M. Head, and A. Ivanov. 2006. “Design Aesthetics Leading to M-Loyalty in Mobile
Commerce.”Information and Management 43 (8): 950–963. doi:10.1016/j.im.2006.08.009.
Cyr, D., M. Head, H. Larios, and P. Bing. 2009. “Exploring Human Images in Website Design: A
Multi-Method Approach.”MIS Quarterly 33 (3): 539–566.
Davis, F., R. P. Bagozzi, and P. R. Warshaw. 1992. “Extrinsic and Intrinsic Motivation to Use
Computers in the Workplace.”Journal of Applied Social Psychology 22 (14): 1111–1132.
doi:10.1111/j.1559-1816.1992.tb00945.x.
602 Q. Wang et al.
Downloaded by [Zhejiang University] at 02:10 13 July 2014
De Angeli, A., M. Matera, M. F. Costabile, F. Garzotto, and P. Paolini. 2003. “On the Advantages of
a Systematic Inspection for Evaluating Hypermedia Usability.”International Journal of Human-
Computer Interaction 15 (3): 315–335. doi:10.1207/S15327590IJHC1503_01.
Deng, L., and M. S. Poole. 2010. “Affect in Web Interfaces: A Study of the Impacts of Web Page
Visual Complexity and Order.”MIS Quarterly 34 (4): 711–730.
Djamasbi, S., M. Siegel, and T. Tullis. 2010. “Generation Y, Web Design, and Eye Tracking.”
International Journal of Human-Computer Studies 68 (5): 307–323. doi:10.1016/j.
ijhcs.2009.12.006.
Djamasbi, S., M. Siegel, T. Tullis, and R. Dai. 2010. Efficiency, Trust, and Visual Appeal: Usability
Testing through Eye Tracking. System Sciences (HICSS),2010 43rd Hawaii International
Conference.
Éthier, J., P. Hadaya, J. Talbot, and J. Cadieux. 2006. “B2C Web Site Quality and Emotions during
Online Shopping Episodes: An Empirical Study.”Information & Management 43: 627–639.
doi:10.1016/j.im.2006.03.004.
Éthier, J., P. Hadaya, J. Talbot, and J. Cadieux. 2008. “Interface Design and Emotions Experienced
on B2C Web Sites: Empirical Testing of a Research Model.”Computers in Human Behavior 24:
2771–2791. doi:10.1016/j.chb.2008.04.004.
Fornell, C., and D. F. Larcker 1981. “Evaluating Structural Equation Models with Unobservable
Variables and Measurement Error.”Journal of Marketing Research 18 (1): 39–50.
Garrett, J. J. 2008. The Elements of User Experience: User-Centered Design for the Web. Beijing:
China Machine Press.
Gefen, D., D. Straub, and M. C. Boudreau. 2000. “Structural Equation Modeling and Regression:
Guidelines for Research Practice.”Communications of the Association for Information Systems
4 (7): 1–70.
Gefen, D., and D. W. Straub. 2003. “Managing User Trust in B2C e-Services.”E-Service Journal 2
(2): 7–24.
Gefen, D., and D. W. Straub. 2004. “Consumer Trust in B2C E-Commerce and the Importance of
Social Presence: Experiments in E-Products and E-Services.”Omega 32 (6): 407–424.
doi:10.1016/j.omega.2004.01.006.
Hassanein, K., and M. Head. 2006. “The Impact of Infusing Social Presence in the Web Interface:
An Investigation across Product Types.”International Journal of Electronic Commerce 10 (2):
31–55. doi:10.2753/JEC1086-4415100202.
Hassanein, K., and M. Head. 2007. “Manipulating Perceived Social Presence through the Web
Interface and Its Impact on Attitude Towards Online Shopping.”International Journal of
Human –Computer Studies 65 (8): 689–708. doi:10.1016/j.ijhcs.2006.11.018.
Henseler, J., and G. Fassott. 2010. “Testing Moderating Effects in PLS Path Models: An Illustration
of Available Procedures.”In Handbook of Partial Least Squares, edited by V. Esposito Vinzi,
W. W. Chin, J. Henseler, and H. Wang, 713–735. Berlin: Springer-Verlag. doi:10.1007/978-3-
540-32827-8_31.
Hess, E. H., and J. M. Polt. 1960. “Pupil Size as Related to Interest Value of Visual Stimuli.”
Science 132 (3423): 349–350. doi:10.1126/science.132.3423.349.
Holbrook, M. B., and R. Batra. 1987. “Assessing the Role of Emotions as Mediators of Consumer
Responses to Advertising.”The Journal of Consumer Research 14 (3):. doi:10.1086/209123.
Hsieh, Y. C., H. C. Chiu, and M. Y. Chiang. 2005. “Maintaining a Committed Online Customer: A
Study across Search-Experience-Credence Products.”Journal of Retailing 81 (1): 75–82.
doi:10.1016/j.jretai.2005.01.006.
Hsu, M. H., and C. M. Chiu. 2004. “Predicting Electronic Service Continuance with a Decomposed
Theory of Planned Behavior.”Behavior and Information Technology 23 (5): 359–373.
doi:10.1080/01449290410001669969.
Huang, M. H. 2003. “Designing Website Attributes to Induce Experiential Encounters.”Computers
in Human Behavior 19: 425–442. doi:10.1016/S0747-5632(02)00080-8.
Hwang, Y., and D. J. Kim. 2007. “Customer Self-Service Systems: The Effects of Perceived Web
Quality with Service Contents on Enjoyment, Anxiety, and E-Trust.”Decision Support Systems
43 (3): 746–760. doi:10.1016/j.dss.2006.12.008.
Izard, C. E. 1992. “Basic Emotions, Relations among Emotions, and Emotion–Cognition Relations.”
Psychological Review 99 (3): 561–565. doi:10.1037/0033-295X.99.3.561.
Johnson, A. R., and D. W. Stewart. 2005. “A Reappraisal of the Role of Emotion in Consumer
Behavior: Traditional and Contemporary Approaches.”In Review of Marketing Research
Enterprise Information Systems 603
Downloaded by [Zhejiang University] at 02:10 13 July 2014
(Review of Marketing Research, Volume 1), edited by N. K. Malhotra, 3–34. Emerald Group
Publishing Limited.
Karanika, M. 2007. Visual Perception: A Cognitive Process. http://www.docin.com/p-428396446.
html, 2007-11-26/2011-04-20.
Karatekin, C. 2007. “Eye Tracking Studies of Normative and Atypical Development.”
Developmental Review 27 (3): 283–348. doi:10.1016/j.dr.2007.06.006.
Lavie, T., and N. Tractinsky. 2004. “Assessing Dimensions of Perceived Visual Aesthetics of Web
Sites.”International Journal of Human –Computer Studies 60 (3): 269–298. doi:10.1016/j.
ijhcs.2003.09.002.
Lazarus, R. S. 1982. “Thoughts on the Relations between Emotion and Cognition.”American
Psychologist 37: 1019–1024. doi:10.1037/0003-066X.37.9.1019.
Lin, A., S. Gregor, and M. Ewing. 2008. “Developing a Scale to Measure the Enjoyment of Web
Experiences.”Journal of Interactive Marketing 22 (4): 40–57. doi:10.1002/dir.20120.
Lin, A., and O. Vasilyeva, 2011. Envisioning the Concept of Emotions for Theory Development and
Testing in Information Systems Research: A Study of One Positive Emotion –Enjoyment,
Proceedings of JAIS Theory Development Workshop. Sprouts: Working Papers on Information
Systems, 11 (155). http://sprouts.aisnet.org/11-155.
Lin, C. P., and A. Bhattacherjee. 2010. “Extending Technology Usage Models to Interactive
Hedonic Technologies: A Theoretical Model and Empirical Test.”Information Systems
Journal 20 (2): 163–181. doi:10.1111/j.1365-2575.2007.00265.x.
Lin, S. 2008. Structural Equation Modeling Theory and the Application of AMOS. Wuhan: Central
China Normal University Press.
Lindgaard, G., G. Fernandes, C. Dudek, and J. Brown. 2006. “Attention Web Designers: You Have
50 Milliseconds to Make a Good First Impression!.”Behaviour and Information Technology 25
(2): 115–126. doi:10.1080/01449290500330448.
Lopatovska, I., and I. Arapakis. 2011. “Theories, Methods and Current Research on Emotions in
Library and Information Science, Information Retrieval and Human–Computer Interaction.”
Information Processing and Management 47 (4): 575–592. doi:10.1016/j.ipm.2010.09.001.
Machleit, K. A., and S. P. Mantel. 2001. “Emotional Response and Shopping Satisfaction:
Moderating Effects of Shopper Attributions.”Journal of Business Research 54 (2): 97–106.
doi:10.1016/S0148-2963(99)00093-4.
Mehrabian, A., and J. A. Russell. 1974. An Approach to Environmental Psychology. Cambridge:
The MIT Press.
Menon, S., and B. Kahn. 2002. “Cross-Category Effects of Induced Arousal and Pleasure on the
Internet Shopping Experience.”Journal of Retailing 78 (1): 31–40. doi:10.1016/S0022-4359
(01)00064-1.
Mummalaneni, V. 2005. “An Empirical Investigation of Web Site Characteristics, Consumer
Emotional States and On-Line Shopping Behaviors.”Journal of Business Research 58 (4):
526–532. doi:10.1016/S0148-2963(03)00143-7.
Nelson, P. 1974. “Advertising as Information.”Journal of Political Economy 82 (4): 729–754.
Parboteeah, D. V., J. S. Valacich, and J. D. Wells. 2009. “The Influence of Website Characteristics
on a Consumer’s Urge to Buy Impulsively.”Information Systems Research 20 (1): 60–78.
doi:10.1287/isre.1070.0157.
Partala, T., and V. Surakka. 2003. “Pupil Size Variation as an Indication of Affective Processing.”
International Journal of Human–Computer Studies 59 (1–2): 185–198. doi:10.1016/S1071-
5819(03)00017-X.
Pavlou, P. A., and M. Fygenson. 2006. “Understanding and Predicting Electronic Commerce
Adoption: An Extension of the Theory of Planned Behavior.”MIS Quarterly 30 (1): 115–143.
Plutchik, R. 1982. “A Psychoevolutionary Theory of Emotions.”Social Science Information 21 (4–
5): 529–553. doi:10.1177/053901882021004003.
Qiu, L., and I. Benbasat. 2009. “Evaluating Anthropomorphic Product Recommendation Agents: A
Social Relationship Perspective to Designing Information Systems.”Journal of Management
Information Systems 25 (4): 145–182. doi:10.2753/MIS0742-1222250405.
Richard, M. O. 2005. “Modeling the Impact of Internet Atmospherics on Surfer Behavior.”Journal
of Business Research 58 (12): 1632–1642. doi:10.1016/j.jbusres.2004.07.009.
Rivard, S., and S. L. Huff. 1988. “Factors of Success for End-User Computing.”Communications of
the ACM 31 (5): 552–561. doi:10.1145/42411.42418.
604 Q. Wang et al.
Downloaded by [Zhejiang University] at 02:10 13 July 2014
Seo, Y. W., S. W. Chae, and K. C. Lee 2012. “The Impact of Human Brand Image Appeal on Visual
Attention and Purchase Intentions at an E-Commerce Website.”In ACIIDS 2012, Part III, LNAI
7198, edited by J.-S. Pan, S.-M. Chen, and N. T. Nguyen, 1–9. Berlin: Springer-Verlag.
Sherman, E., A. Mathur, and R. Smith. 1997. “Store Environment and Consumer Purchase
Behavior: Mediating Role of Consumer Emotions.”Psychology and Marketing 14 (4): 361–
378. doi:10.1002/(SICI)1520-6793(199707)14:4<361::AID-MAR4>3.0.CO;2-7.
Shiv, B., G. Loewenstein, A. Bechara, H. Damasio, and A. R. Damasio 2005. “Investment Behavior
and the Negative Side of Emotion.”Psychological Science 16 (6): 435–439.
Straub, D., and E. Karahanna. 1998. “Knowledge Worker Communications and Recipient
Availability: Toward a Task Closure Explanation of Media Choice.”Organization Science
9 (2): 160–175. doi:10.1287/orsc.9.2.160.
Sun, H., and P. Zhang. 2006. “The Role of Affect in IS Research: A Critical Survey and A Research
Model.”In HCI in MIS (I): Foundations, edited by P. Zhang and D. Galletta, Series of
Advances in Management Information Systems. Armonk, NY: M.E. Sharpe Publisher.
Swan, K., and L. Shih. 2005. “On the Nature and Development of Social Presence in Online Course
Discussions.”Journal of Asynchronous Learning Networks 9 (3): 115–136.
Turley, L. W., and R. E. Milliman. 2000. “Atmospheric Effects on Shopping Behavior.”Journal of
Business Research 49 (2): 193–211. doi:10.1016/S0148-2963(99)00010-7.
Urry, H. L., C. M. van Reekum, T. Johnstone, and R. J. Davidson. 2009. “Individual Differences in
Some (But Not All) Medial Prefrontal Regions Reflect Cognitive Demand while Regulating
Unpleasant Emotion.”Neuroimage 47 (3): 852–863. doi:10.1016/j.neuroimage.2009.05.069.
Van der Heijden, H. 2003. “Factors Influencing the Usage of Websites: The Case of a Generic Portal
in the Netherlands.”Information and Management 40 (6): 541–549. doi:10.1016/S0378-7206
(02)00079-4.
Van der Heijen, H. 2004. “User Acceptance of Hedonic Information System.”MIS Quarterly 28 (4):
695–704.
Vertegaal, R., and Y. Ding, 2002. Explaining Effects of Eye Gaze on Mediated Group
Conversations: Amount or Synchronization? In ACM Conference on Computer Supported
Cooperative Work, New Orleans, LA, 41–48. New York: ACM Press.
Wang, Y. J., M. S. Minor, and J. Wei. 2011. “Aesthetics and the Online Shopping Environment:
Understanding Consumer Responses.”Journal of Retailing 87 (1): 46–58. doi:10.1016/j.
jretai.2010.09.002.
Watson, L., and M. T. Spence. 2007. “Causes and Consequences of Emotions on Consumer
Behaviour: A Review and Integrative Cognitive Appraisal Theory.”European Journal of
Marketing 41 (5/6): 487–511. doi:10.1108/03090560710737570.
Weathers, D., S. Sharma, and S. L. Wood. 2007. “Effects of Online Communication Practices on
Consumer Perceptions of Performance Uncertainty for Search and Experience Goods.”Journal
of Retailing 83 (4): 393–401. doi:10.1016/j.jretai.2007.03.009.
Wu, J. J., and Y. S. Chang. 2005. “Towards Understanding Members’Interactivity, Trust, and Flow
in Online Travel Community.”Industrial Management & Data Systems 105 (7): 937–954.
doi:10.1108/02635570510616120.
Yalch, R. F., and E. R. Spangenberg. 2000. “The Effects of Music in a Retail Setting on Real and
Perceived Shopping Times.”Journal of Business Research 49: 139–147. doi:10.1016/S0148-
2963(99)00003-X.
Enterprise Information Systems 605
Downloaded by [Zhejiang University] at 02:10 13 July 2014
