THE INFLUENCE OF WEBSITE QUALITY ON ONLINE PURCHASE

INTENTIONS: UTILITARIAN AND HEDONIC IMPACTS

ABSTRACT

With the growth and penetration of the Internet, an increasing number of consumers

are shopping online. It is important to analyze how customers make online purchase

decision. However, there is little comprehensive model targeting Chinese in online

purchase behavior. Based on S-O-R paradigm, this research posits a model to assess

how website quality and shopping value influence purchase intention via the mediator

of satisfaction and trust. Using an online survey of 427 Chinese adult respondents, the

proposed model was tested by Structural Equation Modeling. The findings indicate

that under the positive impact of shopping value (utilitarian value and hedonic value),

only two dimensions in website quality (information quality and privacy/security)

positively influences satisfaction and trust, and then purchase intention. Visual appeal

fails to show such a relationship. The study sheds light on online purchase behavior

research in China as well as e-commerce marketing strategy for online sellers.

Keywords: purchase behavior, e-commerce, website quality, utilitarian value, hedonic

value

THE INFLUENCE OF WEBSITE QUALITY ON ONLINE PURCHASE

INTENTIONS: UTILITARIAN AND HEDONIC IMPACTS

1
 INTRODUCTION

The Internet has evolved as a vital communication channel, changing people’s

purchase behavior in an ever increasing range of activities . One significant impact of

the Internet is that it is a catalyst for a new consumption pattern, namely e-commerce .

In China, the number of people using the Internet expanded to 688 million, by the end

of 2015 . This included 413 million online consumers in China, with a growth rate of

14.3% compared to the previous year (CNNIC 2016). In 2015, gross e-commerce

merchandise volume reached 16.2 trillion yuan, and was projected to reach 27.5

trillion yuan in 2018 . Online shopping is so convenient and economic that allows

customers to purchase products at any time and any place by providing a spectrum of

online products with lower prices .

One implication of e-commerce is that there is the risk of divergent online

information between buyers and sellers, which is called information asymmetry.

Consumers can only assess product quality based on website quality (i.e., website

signals like product description, third party certification, and return policy which

influence customers’ e-shopping experience) instead of encountering the physical

product to assess its true value . Thus, buyers have potential communication problems

with sellers resulting in the receipt of unsatisfactory products .

Purchase intention is defined as a mental state that reflects the consumer’s decision to

acquire a product or service in the immediate future . Previous research has

2
demonstrated that website quality plays a pivotal role in positively influencing the

perception of product quality and online purchase intention . Website quality is an

external factor influencing customers’ decision making, however, internal factor --

shopping value (utilitarian value and hedonic value) is less to be analyzed together

with website quality.

The objective of this research is to work on a comprehensive study to investigate how

website quality and shopping value exert influences on online purchase intention in

China. The research results can benefit both theoretical and practical aspects. It can

contribute to e-commerce customers’ research by constructing a comprehensive

model from both external and internal factors . By positing a new model, information

academics can gain better understanding of consumers’ online purchase behaviors. In

a more practical way, after understanding customer online purchase behavior, sellers

can: improve their performance; optimize their strategy; and make better use of e-

channels to attract and retain consumers in the intensely competitive online market. In

addition, customers benefit from better products and services to enjoy better online

shopping experience.

Theoretical background focuses on: signalling theory, S-O-R paradigm and purchase

intention. Then the nine hypotheses are described and the initial research model is

outlined. Research methodology, results and discussion are presented later. Finally,

the conclusion, implications and limitations are stated.

3
 Theoretical background

Signalling Theory

Signalling theory can be defined as a framework to describe the behavior between two

parties (individuals or organizations) when they are facing asymmetric information in

pre-contractual contexts . In e-commerce settings, an information asymmetry between

online sellers and buyers potentially occurs when consumers make purchase decisions

on the website signals without encountering the real product, increasing the

probability of inaccurate assessment of product quality .

Website quality acts as website signals to transfer product information from online

sellers to online buyers. Harris and Goode (2010) evaluated website quality from four

aspects: aesthetic appeal; layout; functionality; and financial security. Kim et al.

(2012) identified three dimensions: system; information; and service qualities. Moez

(2013) assessed the effect of website attributes (navigation, visual appearance and

customized preview) on consumers’ commitment to website and buying intention.

Similarly, Liu et al. (2013) revealed that product availability, website ease of use and

visual appeal are three website cues that affect online impulse buying. Ease of use, an

important factor among website cues, appears in many articles . In summary, the

literature indicates that website quality contains information quality, system quality,

service quality, ease of use and visual appeal (see Table 1).

Table 1. Website Quality Dimensions

4
 Author
Harris C. J. Kim
Website quality Online and Kim Moez and
Liu et al.
dimension signals Good et al. (2013 Lenno
(2013)
e (2012 ) n
(2010) ) (2013)
Information Product
√ √
quality description
Layout, page
Visual appeal √ √ √ √
design
Warranties,
third-party
Privacy/
seals, √ √ √
security
financial
security
Return
policy, refund
Service quality √ √ √
policy, quick
response
Search
Ease of use engine, √ √
navigation

S-O-R Paradigm

The S-O-R paradigm, initially forwarded by Mehrabian (1974), describes consumer

behavior into three steps, namely: external stimuli; inner organism changes; and

responsive behavior. Online environmental stimuli are comprised of ambient factors,

design factors and social factors . Organism includes emotion and cognition states

that connect the relationships between the stimulus and the individual’s responses,

such as satisfaction and trust . Response refers to behavioral intention like purchase

intention, repurchase intention, word of mouth and brand loyalty . There is a large

body of knowledge concentrating on analyzing online shopping purchase intention

according to S-O-R .

5
 Purchase Intention

Based on the signalling theory, customers view the quality of products through

website quality (online signals), which acts as external influences on consumer

purchase intention. Previous research found that purchase intention is positive related

with website quality (external factor). Chang and Chen (2008) investigated whether

online environment cues (web site quality and web site brand) affect customer

purchase intention towards an online retailer and whether this impact is mediated by

customer trust and perceived risk. Their findings confirmed that web site quality and

web site brand affect consumers’ trust and perceived risk, and in turn, consumer

purchase intention. Kim and Lennon (2013) used the S-O-R paradigm as a model with

reputation and website quality as stimuli that affect purchase intention. Results

showed that reputation and four website quality dimensions have a significant positive

effect on consumers’ emotions.

Additionally, shopping value, like utilitarian value and hedonic value, which is the

consumers’ internal factor, is positively associated with buyers’ purchase intention .

So there is a need to combine consumers’ external factors (e.g., website quality) and

internal factors (e.g., shopping value) to produce a more comprehensive model of

online shopping behavior. Chih and Lin (2011) investigated online shopping in

Taiwan with a model that included both consumer values (utilitarian value and

hedonic value) and website perspectives to evaluate the influence on consumers’

6
purchase intention. However, the range of factors they applied cannot describe whole

website quality, e.g. lacking visual appeal or ease of use. Kim et al. (2012) examined

the effects of website qualities (system, information and service qualities) on Internet

shopping values (utilitarian and hedonic values) and their influences on satisfaction

and repurchase intention in South Korea. However, the range of quality dimensions

they applied are not exhaustive, e.g. visual appeal can be added to their analysis.

Although there are models consisting from two perspectives, they are still not

comprehensive and needed to be expanded to consider a wider range of potential

issues. Furthermore, e-commerce is expanding in China, but there are few

comprehensive research models targeting the Chinese online market. Therefore, based

on S-O-R paradigm, it is timely to analyze online purchase behavior in China by the

joint influence of website quality and shopping value.

RESEARCH MODEL AND HYPOTHESES DEVELOPMENT

To analyze consumer online purchase behavior, a research model is proposed based

on S-O-R paradigm. S-O-R paradigm shows a causal relationship (stimuliinner

organism changesresponsive behavior) that posits to illustrate the process and

emotional change before purchase decision. In this study, stimuli include website

quality and shopping value. Inner organism changes refer to attitude (satisfaction and

trust). Responsive behavior means purchase intention. The model (see Figure 1)

7
explores the influence of website quality and shopping value on customers’ online

purchase behavior where trust and satisfaction are mediators.

Figure 1. Research Model

Website Quality

Website quality covers information quality, visual appeal, privacy/security, service

quality and ease of use. Mai (2013) defined information quality as a product of the

degree to which the exchange and production of meaning has been successful.

Previous studies demonstrate that consumers feel satisfied with an e-commerce

system if the system offers high quality information . Once customers process useful,

rich and understandable information, they can reach a better understanding of the

product, reducing the difference between expectations and the actual product,

resulting in higher satisfaction for consumers . Furthermore, consumers will generate

trust towards an online product when they obtain an accurate assessment of the

8
product through complete and detailed information. Mai (2013) noted that

information with good quality usually transfers ideas efficiently and effectively,

allowing users to feel it is a trustworthy product. Thus, high information quality is

associated with high user satisfaction as well as trust . Thus, the hypotheses are listed

as follows.

Hypothesis 1a Information quality has a positive influence on satisfaction.

Hypothesis 1b Information quality has a positive influence on trust.

Visual appeal means well-designed format for presenting product information or

layout information, such as the size, the color, and the amount of pictures . Liu et al.

(2013) argued that a visually-appealing website would enhance the felt gratification of

purchase online. Wells et al. (2011) stressed that mediated by user satisfaction,

perceived visual appeal has an indirect influence on impulse purchases. Additionally,

websites with an acceptable visual interface often result in a better online shopping

experience, then generate trust and enhance the relationship with the online sellers . In

summary, effective website design influences customers’ decision on purchase

through positive attitudes like satisfaction and trust . The hypotheses are listed as

follows.

Hypothesis 2a Visual appeal has a positive influence on satisfaction.

Hypothesis 2b Visual appeal has a positive influence on trust.

9
Privacy/Security refers to customers’ perception towards security of payment

procedures and privacy of shared information . Security perceptions have a link to

satisfaction . Customers will be satisfied when they perceive that a website is safe.

Additionally, information security plays an important role in the website reliability

and influences trust on transactions . Many e-commerce websites use third-party

certification to guarantee and assure the trust from consumers . Therefore,

privacy/security has a positive impact on satisfaction and trust before purchase

happens. The hypotheses are listed as follows.

Hypothesis 3a Privacy/Security has a positive influence on satisfaction.

Hypothesis 3b Privacy/Security has a positive influence on trust.

Service quality can be defined as a user’s satisfaction with customer service from

online sellers or website . Petter et al. (2012) noted that users’ perceptions of website

services have a positive effect on satisfaction. If an e-commerce website fails to

provide good service, consumers will be dissatisfied and stop seeking information

there. Additionally, high service quality encourages consumers to form a positive

attitude towards a website, become confident in online sellers’ reputation and,

therefore, generate trust in an online store . Ha and Stoel (2009) also found that

prompt product delivery and ease of product return significantly increase consumers’

trust in online website. Accordingly, good online service quality generates positive

10
perspectives from consumers, resulting in satisfaction and trust. Thus, the hypotheses

are listed as follows.

Hypothesis 4a Service quality has a positive influence on satisfaction.

Hypothesis 4b Service quality has a positive influence on trust.

Ease of use is selected from the TAM model, which suggests that individuals accept

information technology if they believe in its positive performance . It can be defined

as customers’ perception towards finishing a task without much effort and how easy

the online operation is with the help of navigation and a search engine . Prior studies

suggest that if a website is easy to use, it is possible to raise consumers' positive

emotions like satisfaction, therefore encouraging purchase behavior . The perceived

ease of use can also increase trust in the online store when online shoppers recognize

that the online store design the system from the users’ perspective. Conversely, if a

website is difficult to operate, users might doubt the professionalism of the website,

thereby decreasing their perception of trust . Thus, the research hypothesis is as

follows.

Hypothesis 5a Ease of use has a positive influence on satisfaction.

Hypothesis 5b Ease of use has a positive influence on trust.

Shopping Value

11
Shopping value (utilitarian value and hedonic value) represents consumer’s

motivation to shop online . Customers with a stronger utilitarian value seek product

information online based on an inner need to efficiently acquire merchandise through

task-oriented shopping activities . Task-oriented shopping is the main theme in

utilitarian value, so that customers purchase products based on clear goals and

efficient approaches, with the hope of convenience, monetary savings, quick response

time, usability and rich product information . Kim et al. (2012) argued that a

customer’s perception of utilitarian value is a key determinant of customer

satisfaction and loyalty, so satisfied customers are more likely to purchase. Kuan et al.

(2008) noted that utilitarian values are highly related to the likelihood of customer

satisfaction and purchase. Additionally, utilitarian value has an indirect influence on

purchase intention that is moderated by consumers’ trust of the product . A potential

consumer's positive perception of website usability enhances the source’s credibility

and thereby trustworthiness . Thus, utilitarian value has a positive link with

satisfaction and trust. The hypotheses are listed as follows.

Hypothesis 6a Utilitarian value has a positive influence on satisfaction.

Hypothesis 6b Utilitarian value has a positive influence on trust.

Instead of goal-oriented need, hedonic value tends to be more subjective and is

usually based on: entertainment; pleasure; enjoyment; experience; and fun . Avcilar

and Ozsoy (2015) reported that hedonic value is closely related to shopping

12
satisfaction and has a significant effect on consumer online shopping intentions.

Because hedonic pleasure enhances an online shopper’s emotional benefit and

satisfaction toward a website, customers are willing to spend more time seeking other

items online . Customers will also build sense of trust and tend to browse websites

more often since they enjoy the process of surfing online . Therefore, hedonic value is

highly linked to satisfaction and trust. The hypotheses are listed as follows.

Hypothesis 7a Hedonic value has a positive influence on satisfaction.

Hypothesis 7b Hedonic value has a positive influence on trust.

Attitude

Satisfaction is defined as the summary psychological state and affective response to

the overall online information seeking experience . Many researchers have pointed out

the positive relationship between satisfaction and purchase intention . Shiau and Luo

(2012) illustrated that customer satisfaction can strengthen the relationship between

customers and online sellers, resulting in a purchase trend. Similarly, Avcilar and

Ozsoy (2015) mentioned that shopping satisfaction has a significant effect on

consumer online shopping intentions.. Thus, the degree of satisfaction is highly linked

to purchase intention . The hypothesis is listed as follows.

Hypothesis 8 Satisfaction has a positive influence on purchase intention.

13
Trust refers to subjective guarantees that the online store brand can meet its

obligations, will behave as promised, and genuinely cares about customers . Many e-

commerce studies conclude that trust has a positive impact on the behavioral

intentions of online consumers . The extent to which consumers trust the website is

linked to further purchase, repurchase or even loyalty . Thus, trust, considered as an

influencer, is an important influence on the final online purchase decision .The

hypotheses are listed as follows.

Hypothesis 9 Trust has a positive influence on purchase intention.

METHODOLOGY

Data Collection and Sampling

After pilot study, the final survey was distributed via Baidu Tieba, the largest Chinese

online community. People aged over 18 years old from China who had made online

purchases during the last 12 months were invited to complete the survey in an

anonymous way. According to Yamane , when the size of population is larger than

100,000, sample size is required to be 400 with 95% confidence level and ±5%

Precision Levels. This study received 427 samples: two respondents refused to

participate in the study and six had not shopped online in the 12 months, so were

excluded. As Table 2 shows, the majority of participants (63.46%) were aged between

18-30 and the frequency of females (N=210) and males (N=217) were nearly

identical. Students (29.27%) are the largest proportion in the sample, followed by

14
people working in private firms (21.78%). In terms of education and income, more

than half of respondents (56.67%) hold a bachelor degree and nearly half (48.34%)

earned 3,000-5,000RMB per month.

Table 2. Subjects’ Demographic Information

Item Category Frequency Percentage

18-24 132 30.91%
25-30 139 32.55%
Age
31-35 88 20.61%
>35 68 15.93%
Female 210 54.78%
Gender
Male 217 45.22%
High school 14 3.28%
Associate 66 15.46%
Education background Bachelor 242 56.67%
Master 98 22.95%
PHD 7 1.64%
Student 125 29.27%
Self-employed 41 9.60%
Private enterprise 93 21.78%
Foreign company 32 7.49%
Occupation
State-owned enterprise 61 14.29%
Public institution 50 11.71%
Civil servant 7 1.64%
Others 18 4.22%
Less than 3000 24 7.95%
3000 - 5000 146 48.34%
Income 5000-8000 85 28.15%
8000-12000 15 4.97%
more than 12000 32 10.60%
Note: n=419

Measurement development

15
A seven-point Likert Scale (1 = strongly disagree, 4 = neither agree nor disagree, and

7 = strongly agree) was used to measure these construct items. The scale items for

information quality were adapted from Mathwick et al. (2002) and Wolfinbarger and

Gilly (2003). Items for visual appeal were derived from Mathwick et al. (2002).

Privacy/security was taken from scales provided by Wolfinbarger and Gilly (2003)

and Cheung and Lee (2001). Service quality items were adapted from Parasuraman et

al. (2005) and Gefen (2002). Perceived ease of use was taken from TAM model by

Davis (1989) and Devaraj et al. (2002). Items for utilitarian and hedonic value were

both chosen from Overby and Lee (2006), Bridges and Florsheim (2008) and Gupta

and Kim (2010). Satisfaction items were adapted from Spreng et al. (1996) and trust

items were derived from Gefen et al. (2003). The source of purchase intention was

Loiacono et al. (2007) and Srinivasan et al. (2002).

Data Analysis

In the process of quantitative data analysis, the initial step was to do Exploratory

Factor Analysis (EFA) in SPSS, which is a statistical approach for exploring the

factor structure by determining the correlation among the variables in a dataset. The

second step is to test Confirmatory Factor Analysis (CFA) in SPSS and AMOS,

which is to confirm the factor structure extracted in the EFA. Model fit assessment

refers to how well the proposed model accounts for correlations between variables in

the dataset test. Reliability and validity of the data are then evaluated in the

measurement model. Reliability can ensure that measurement items have internal

16
consistency, which describe the same construct . Construct validity determines

whether items truly measure the design of the research, including convergent validity

and discriminant validity . Convergent validity shows how well the items are designed

to measure certain construct convergence . Discriminant validity is used to confirm

that the measurement instrument is different from other similar constructs .

The third step is to test the hypotheses by structural equation modeling (SEM) in

AMOS. Steenkamp and Baumgartner (2000) noted two key benefits of this method.

First, SEM is able to incorporate measurement error explicitly into models and

analyze its influence on the degree of fit. Second, compared to multiple regressions,

the total casual relations inside model can be studied at one click without additional

calculation steps. Additionally, mediation test is used to provide a more accurate

explanation for the causal effect that the antecedent has on the dependent variable.

RESULT

Shopping Behavior Analysis

The results of shopping behavior analysis are illustrated in Table 3. Fashion products

remained the most popular category at 43.68%, while electronics and food follow

with only 15.27% and 13.13% respectively. Participants showed a trend for e-

shopping: approximately half made online purchase more than three times a month.

Taobao received the most votes (51.79%) and Tmall came second (24.58%), both of

which belong to Alibaba. As Taobao is the largest e-commerce website in China, it

17
owns the biggest market share . JD came third as it is the third largest Chinese e-

commerce company . In terms of online payment method, 81.62% of participants

chose to use third party payment, followed by credit card (10.5%) and debit card

(5.73%).

Table 3. Online Shopping Experience

Item Category Freq. Percentage

Electronics (e.g., phones, cameras &
64 15.27%
computers)
Fashion 183 43.68%
Food 55 13.13%
Books & CD 26 6.21%
Most frequently Medicine 3 0.72%
bought items Travel (e.g., accommodation & flight
17 4.06%
tickets)
Cosmetics 15 3.58%
Household items (e.g., home
44 10.50%
decoration, furniture & appliances)
Others 12 2.86%
More than three times a month 105 45.65%
Once or twice a month 84 36.52%
How often do you
Once or twice a quarter 32 13.91%
shop online?
Once half a year 8 3.48%
Once a year 1 0.43%
Taobao.com 217 51.79%
Tmall.com 103 24.58%
Kaola.com 7 1.67%
Favorite online
Jd.com 60 14.32%
shopping site
Amazon.cn 13 3.10%
Dangdang.com 5 1.19%
others 14 3.34%
Third party payment (e.g., AliPay) 342 81.62%
Most used online Credit card 44 10.50%
payment method Debit card 24 5.73%
cash on delivery 9 2.15%
Note: n=419

18
 Exploratory Factor Analysis

SPSS was employed to conduct EFA data analysis. KMO was first tested, where the

value is below 0.5 it is not suitable for factor analysis. KMO value is totally above 0.7

(see Table 4), showing that all the data are suitable for factor analysis. To test

construct validity, principle component analysis with varimax rotation is adopted

where the coefficient display is sorted by size and absolute value below 0.5 is

suppressed. All measurement items load together to the corresponding constructs (see

Appendix). The factor loadings are all above 0.65, showing good validity .

Table 4. KMO Value

Bartlett's Test of
KMO Measure of Sphericity
Sampling Adequacy. Approx. Chi- Sig
df
Square .
15
Website quality 0.911 3791.947 0
3
Shopping value 0.784 1238.103 28 0
Trust &
Satisfaction 0.916 2631.389 55 0
purchase intention

Confirmatory Factor Analysis

In the assessment of model fit, all fit indices showed that the model data provided an

acceptable fit (CMIN/DF=2.129; P=0; CFI=0.923; GFI=0.846; AGFI=0.814;

RMSEA=0.055). They are summarized in Table 5 and the theory is supported by .

19
Table 5. Assessment of Model Fit

Fit Indices CMIN/DF P CFI GFI AGFI RMSEA

Result 2.129 0 0.923 0.846 0.814 0.055
Criteria <3.0 significant >0.9 >0.8 >0.8 <0.07
Note: CMIN: Chi-square; DF: degrees of freedom; CFI: comparative fit index;
GFI: goodness of fit index; AGFI: adjusted goodness of fit index; RMSEA: root
mean square error of approximation

Reliability analysis is used to ensure that the measurement items describe the same

construct. Both value of Cronbach’s Alpha and composite reliability (CR) in each

construct is above 0.7 and Corrected Item-Total Correlation in each indicator is above

0.6 (see Table 6), representing a good performance of internal consistency reliability .

Table 6. Reliability of Construct

Corrected Cronbach's
Cronbach's
Construct Indicators Item-Total Alpha if Item CR
Alpha
Correlation Deleted
INFOQ1 0.652 0.82
Information INFOQ2 0.749 0.778
0.848 0.851
quality INFOQ3 0.709 0.798
INFOQ4 0.639 0.829
VA1 0.757 0.846
Visual
VA2 0.772 0.833 0.882 0.882
appeal
VA3 0.788 0.82
Continued

Corrected Cronbach's
Cronbach's
Construct Indicators Item-Total Alpha if Item CR
Alpha
Correlation Deleted
P/S1 0.714 0.796
Privacy/security P/S2 0.778 0.731 0.85 0.854
P/S3 0.675 0.833
20
 SQ1 0.631 0.798
SQ2 0.707 0.764
Service quality 0.83 0.835
SQ3 0.624 0.806
SQ4 0.682 0.776
EOU1 0.632 0.847
EOU2 0.759 0.793
Ease of use 0.857 0.86
EOU3 0.707 0.815
EOU4 0.71 0.814
UV1 0.608 0.767
UV2 0.658 0.741
Utilitarian value 0.798 0.786
UV3 0.735 0.696
UV4 0.705 0.71
HV1 0.664 0.837
HV2 0.743 0.801
Hedonic value 0.856 0.861
HV3 0.756 0.796
HV4 0.71 0.812
SAT1 0.649 0.822
Satisfaction SAT2 0.768 0.71 0.839 0.849
SAT3 0.739 0.741
TR1 0.682 0.832
TR2 0.709 0.82
Trust 0.859 0.862
TR3 0.706 0.826
TR4 0.745 0.803
PI1 0.738 0.866
Purchase PI2 0.708 0.874
0.889 0.891
intention PI3 0.8 0.84
PI4 0.784 0.846

Values of AVE are all above 0.5 (see Table 7), representing a good convergent

validity . Also, the square root of AVE values (shown in the diagonals) are all larger

than the inter-construct correlations, which means it has good discriminant validity .

Table 7. Validation of Measurement Model

AVE TR INFOQ VA PS SQ EOU UV HV SAT PI

21
TR 0.610 0.781
INFOQ 0.590 0.640 0.768
VA 0.715 0.600 0.652 0.845
PS 0.661 0.742 0.603 0.640 0.813
SQ 0.558 0.684 0.625 0.603 0.611 0.747
EOU 0.606 0.467 0.528 0.584 0.527 0.613 0.778
UV 0.540 0.490 0.553 0.550 0.494 0.546 0.675 0.735
HV 0.613 0.456 0.264 0.420 0.381 0.339 0.169 0.311 0.783
SAT 0.653 0.757 0.569 0.638 0.629 0.584 0.598 0.628 0.603 0.808
0.82
PI 0.673 0.772 0.527 0.590 0.605 0.603 0.686 0.680 0.352 0.747 0
Note: AVE is calculated as ρvc(η)=(Σλi2)/[(Σλi2)+Σ(1-λi2)] where λi refers to
the ith standardized component loading and (1-λi2) refers to the ith error
variance; off-diagonals are shared variance; diagonals are square root of AVE
values; TR: trust; INFOQ: information quality; VA: visual appeal; PS:
privacy/security; SQ: service quality; EOU: ease of use; UV: utilitarian value;
HV: hedonic value; SAT: satisfaction; PI: purchase intention

Testing the Hypothesis

After confirming that the model was reliable and valid, Structural Equation Modeling

(SEM) was applied to analyze the causal relationship between variables in the model

(see Table 8). Standardized regression coefficients (β) show the predictive strength of

independent variables and adjusted R-square represents the explanatory power of

constructs in the model .

In the model block 1, which measures the relationship between satisfaction and

website quality as well as shopping value, the adjusted R-square is 0.559, showing

moderate explanatory power . In website quality dimensions, ease of use (β=.216,

p<.001) is found to be the best predictor with high significance and positive

association. Standardized coefficients of privacy/security and information quality are

0.154 and 0.132 respectively, both at a significant level. However, visual appeal
22
(β=.115, p>.01) and service quality (β=.063, p>.01) fail to show a significant positive

relationship with satisfaction. For shopping value, both utilitarian value (β=.131,

p<.01) and hedonic value (β=.277, p<.001) show significant positive associations with

satisfaction.

Model block 2 describes the causal relationship between trust and different

antecedents. The result differed from that in model block 1. Based on adjusted R-

Square, 53.1% of the variance in trust can be explained, which shows a moderate

explanatory power . Information quality (β=.188, p<.001), privacy/security (β=.339,

p<.001), service quality (β=.247, p<.001) and hedonic value (β=.173, p<.001) exert

positive influence on trust. The other three factors fail to show significant association

with trust.

Purchase intention in model block 3 can be explained moderately as the adjusted R-

Square is 0.553 . It shows strong positive association with satisfaction (β=.417, p

<.001) and trust (β=.401, p<.001).

To summarize, the hypotheses are mostly supported by testing. However, visual

appeal is not positive with both satisfaction and trust. Service quality is not

significantly associated with satisfaction. Ease of use and utilitarian value have no

significant influence on trust.

Table 8. Hypotheses and Results

23
Mode Standardize Adjuste
t-
l Hypotheses d Result d R-
value
block Coefficients Square
Information quality
0.132** 2.872 Supported
Satisfaction
Unsupporte
Visual appeal Satisfaction 0.115 2.383
d
Privacy/Security
0.154*** 3.389 Supported
Satisfaction
Service quality Unsupporte 0.567
1 0.063 1.369
Satisfaction d
Ease of use Satisfaction 0.216*** 4.328 Supported
Utilitarian value
0.131** 2.781 Supported
Satisfaction
Hedonic value
0.277*** 7.403 Supported
Satisfaction
Information quality Trust 0.188*** 3.956 Supported
Unsupporte
Visual appeal Trust 0.053 1.077
d
Privacy/Security Trust 0.339*** 7.228 Supported
Service quality Trust 0.247*** 5.185 Supported
2 0.540
Unsupporte
Ease of use Trust -0.035 -0.679
d
Unsupporte
Utilitarian value Trust 0.002 0.041
d
Hedonic value Trust 0.173*** 4.475 Supported
Satisfaction Purchase
0.417*** 9.074 Supported
3 intention 0.553
Trust Purchase intention 0.401*** 8.73 Supported
Note: ***p<.001;**p<.01

Additionally, mediation effect in AMOS was used to test the influence of mediators

between independent variables and dependent variables. Satisfaction and trust exert

the role of mediator in this model. Previous approach was to test the model with

mediator and without the mediator, where there are three types: partial mediation; full

mediation; and indirect mediation . However, recent studies suggest that if the indirect
24
effect is significant, then mediation exists, regardless of whether the direct effect was

significant before adding the mediator . So, there is no need to test the model without

mediator, and mediation is determined purely by the bootstrapped indirect effect.

Bootstrapping involves repeatedly randomly sampling observations with replacement

from the data set to compute the desired statistic in each resample .

Table 9 shows that satisfaction has no mediation effect on the relationship between

service quality and purchase intention, which is obvious and consistent because

service quality has no significant positive influence on satisfaction (see Table 8). All

the other relationships showed a mediation effect by satisfaction. Trust mediates the

causal relationships between information quality, privacy/security, as well as hedonic

value, and purchase intention. The other 3 relationships were not affected.

Table 9. Indirect Influence

Satisfaction as a Mediator Trust as a Mediator

Indirect 2-tailed Indirect 2-tailed
Result Result
significance significance
INFOQ PI 0.004 Mediation 0 Mediation
No
VA PI 0.011 Mediation 0.279
mediation
PS PI 0.003 Mediation 0 Mediation
25
 No
SQ PI 0.145 0 Mediation
mediation
No
EOU PI 0 Mediation 0.521
mediation
No
UV PI 0.003 Mediation 0.978
mediation
HV PI 0 Mediation 0 Mediation
Note: INFOQ: information quality; VA: visual appeal; PS: privacy/security; SQ:
service quality; EOU: ease of use; UV: utilitarian value; HV: hedonic value; PI:
purchase intention

In summary, some hypotheses fail to exist based on the insignificance of casual

relationships. Visual appeal has no significant positive impact on satisfaction and

trust. Service quality fails to show positive relationship with satisfaction and

utilitarian value fails to show positive connection with trust. The proposed model is

modified based on results above in Figure 2.

Figure 2. Modified Model

26
 DISCUSSION

This study examined the effects of website quality and shopping value on purchase

intention via the mediator of satisfaction and trust. These variables were analyzed by

survey in terms of Likert scale measurement items. The findings show that website

quality and shopping value have a positive influence on purchase intention.

Based on the hypotheses testing, information quality and privacy/security mediated by

trust and satisfaction exert a significantly positive influence on customer purchase

intention. This positive relationship can also be demonstrated by the mediation effect

in Table 19-20.

Visual appeal fails to demonstrate a positive association with both trust and

satisfaction, showing a weak influence on purchase intention, which means that

visual appeal is not a strong factor in the customer purchase decision making process.

Ease of use exerts significant influence on satisfaction, but is not positively related to

trust. This means that users will not generate trust for the website or online product

when they encounter a user friendly system. They are aware that a good quality

system does not necessarily equal to a good quality product, which is their real need.

But they will be satisfied with the website since it makes it easier to operate the

system. Service quality is positive associated with trust, but shows poor relationship

to satisfaction. Before purchase, buyers will trust sellers who provide help and solve

their problems. But buyers will not feel satisfied instantly and their satisfactory will

27
only occur when the complete transaction is finished and the product’s quality

matches what buyers have promised online. So buyer satisfaction is hard to gain by

service quality alone. Even so, ease of use and service quality still pose a moderate

influence on purchase intention.

System quality, information quality and service quality are three factors in the IS

success model . Kim et al. (2012) selected them as website quality dimensions into

their model, where system quality combined the meaning of privacy/security and ease

of use. Their findings are consistent with this study’s result that information quality,

system quality, and service quality are positive associated with purchase intention.

In addition, shopping value has a positive impact on trust as well as satisfaction, and

then purchase intention. However, utilitarian value fails to show a positive

relationship to trust. People with utilitarian value desire to purchase a high quality

product. They are goal-oriented, so they will not easily trust online sellers until they

obtain the products that they like . Although utilitarian value doesn’t show a positive

association with trust, other three relationships perform quite well. Overall, shopping

value positively influences purchase intention .

To summarize, a website with good quality, especially accurate product information,

good system quality and considerate service, can largely generate trust and

satisfaction, and then arouse customers’ buying intentions. Utilitarian value and

28
hedonic value both, to some degree, impact the online purchase decision making

process.

CONCLUSION

This study conducted a survey and constructed an evidence-based model to explore

customer online purchase behavior in China. Due to a wide range of merchandise,

Taobao has emerged as the most-preferred e-commerce website. Fashion products are

the most frequently bought items and two thirds of customers purchase online

products every month.

Mediated by satisfaction and trust, website quality (information quality and

privacy/security) has a positive influence on purchase intention. Trust is the mediator

for the relationship between service quality and the purchase intention. Satisfaction is

the mediator for the relationship between ease of use and the purchase intention. In

contrast, visual appeal has no significant positive relationship to purchase intention.

Additionally, shopping value (utilitarian value and hedonic value) positively

influences purchase intention through satisfaction and trust.

IMPLICATIONS AND LIMITIONS

This study provides implications for academics by positing a comprehensive model to

assess online purchase behavior in China. This model can be regarded as a theoretical

basis for further research investigating e-shopping behavior from the consumer

perspective.
29
In addition, it offers suggestions for practitioners by surfacing the relative importance

of different website quality factors. Information quality, security issues, service

quality, and ease of use are four essential factors that should be addressed by online

sellers. For information quality, they should provide accurate, comprehensive and

easy-to-understand product descriptions to reduce buyers’ confusion towards product

quality. In terms of privacy/security, Taobao offers Alipay (a third-party online

payment platform) to guarantee the security of transactions. However, there is always

potential for information leakage. So security or privacy issues remain a challenge for

e-commerce companies and online sellers. For service quality, online sellers with an

appropriate attitude and manner can increasingly build trust from buyers.

Unsatisfactory service includes: delayed response; no return or refund choice; and

slow delivery. Good service mainly depends on online sellers’ willingness to

recognize the customer perspective. Websites should also provide user-friendly and

easy to use interfaces so users will not have difficulties operating the system, creating

a good impression of the site.

There are several limitations to this investigation. First, this study adopts a

quantitative method rather than mixed methods. Further study could add interviews

and case studies to make it a mixed methods approach. By triangulating each result,

the research conclusion could be more convincing. Second, this study adopts the

consumers’ perceptions of website quality. However, this data is subjective and may

30
result in response bias. Third, the sample size is restricted to four-hundred Chinese

participants. While 400 participants are considered a valid sample size for populations

above 100,000 (Yamane 1967): a larger sample could elicit a more representative

result. Future studies could also draw data from U.S. to compare online purchase

behavior between U.S. and China.

ACKNOWLEDGMENT

This research did not receive any specific grant from funding agencies in the public,

commercial, or not-for-profit sectors. I would like to express gratitude to my

supervisor Mr. Peter Warning for support and guidance for this topic. In our

discussions, my idea of the thesis gradually turned from vague to clear from his

feedback. He also suggested appropriate revisions.

31
REFERENCE

32
 APPENDIX

Measurement Items

Indicator
Factors Measured items Source
s

This website provides

abundant and comprehensive
INFOQ1
information regarding product
functions and quality.

This website produced

Wolfinbarge
accurate information
INFOQ2 r and Gilly
Information regarding product functions
(2003);
quality and quality.
Mathwick et
This website provides easy-
al. (2002)
to-understand information
INFOQ3
regarding product functions
and quality.
This website offers active
INFOQ4 user reviews and ratings of its
products.
The page design (i.e. colors,
VA1 layouts, animations, etc.) is
attractive.
Mathwick et
Visual appeal This site looks professionally
VA2 al. (2002)
designed.
The overall look and feel of
VA3
the site is visually appealing.
This website provides good
protection of personal
P/S1
information(e.g.,credit card Wolfinbarge
number & e-mail address) r and Gilly
Privacy/security This website protects (2003);
P/S2
personal buying information. Cheung and
Internet shopping system Lee (2001)
P/S3 generally offers safe
transactions.
Continued

33
 Indicator
Factors Measured items Source
s
The online seller responds to
SQ1
my queries in a timely manner.
The online seller shows a
SQ2 sincere interest in solving my
problem. Parasuraman
The website provides quick et al. (2005);
Service quality SQ3
delivery (i.e.,3 days or less) Gefen (2002)
The website provides me with
convenient options for
SQ4 returning items according to its
product return policy (return in
7 days)
It is easy to learn to use the
EOU1
website.
It is easy to navigate products
EOU2 using bar menu or search Devaraj et al.
Ease of use engine. (2002); Davis
It is easy to find information I (1989)
EOU3
want.
It is quick and easy to complete
EOU4
a transaction at this website.
I go online shopping because I
UV1 can get the same quality
products at a lower price.
I go online shopping because
online shopping saves time Overby and
UV2
compared to in-person Lee (2006);
shopping. Bridges and
Utilitarian
I go online shopping because I Florsheim
value
can quick access to large (2008);
UV3
volumes of product and service Gupta and
information. Kim (2009)
I go online shopping because I
can have a more comfortable
UV4
and convenient shopping
environment
Hedonic value I feel like an escape when I Overby and
HV1
shopping on this website. Lee (2006);
HV2 I enjoy passing the time by Bridges and

34
 shopping online.
I can relieve stress and feel
HV3
relaxed when I shop online.
Florsheim
I look around at items on the
HV4 (2008); Gupta
Internet just for fun.
and Kim
Continued

Indicator
Factors Measured items Source
s
I am satisfied with the shopping
SAT1 experience I have had on this
website.
Using the online shopping
Satisfactio Spreng et al.
SAT2 website makes me feel very
n (1996)
pleased.
Using the online shopping
SAT3 website makes me feel very
content.
TR1 This website is trustworthy.
TR2 This website is honest.
Most of what this web site says
TR3 Gefen et al.
Trust about its products is true.
(2003)
If this web site makes a claim or
TR4 promise about its product, it is
probably true.
I intend to purchase more goods
PI1 from this website in the next few
years.
I intend to acquire product
PI2 Loiacono et al.
information from this website.
Purchase (2007) ;
I will buy an item from this
intention Srinivasan et
PI3 website if I find something that I
al. (2002)
like.
I will visit this web site when I
PI4 want to buy certain items in near
future.

35
 Factor Analysis

Component
Measurement items of website quality
EOU INFOQ SQ VA P/S
It is easy to navigate products using
0.829
bar menu or search engine.
It is easy to find information I want. 0.799
It is quick and easy to complete a
0.745
transaction at this website.
It is easy to learn to use the website. 0.718
This website produced accurate
information regarding product functions 0.789
and quality.
This website offers active user reviews
0.759
and ratings of its products.
This website provides easy-to-
understand information regarding 0.74
product functions and quality.
This website provides abundant and
comprehensive information regarding 0.716
product functions and quality.
The online seller shows a sincere
0.776
interest in solving my problem.
The website provides me with
convenient options for returning items
0.758
according to its product return policy
(e.g. return in 7 days)
The online seller responds to my
0.725
queries in a timely manner.
The website provides quick delivery
0.688
(i.e.,3 days or less)
This site looks professionally designed. 0.825
The page design (i.e. colors, layouts,
0.821
animations, etc.) is attractive.
The overall look and feel of the site is
0.799
visually appealing.
Continued

36
 Component
Measurement items of website quality
EOU INFOQ SQ VA P/S
This website protects personal buying
0.845
information.
This website provides good protection of
personal information (e.g.,credit card 0.763
number & e-mail address)
Internet shopping system generally offers
0.742
safe transactions.
Note: Extraction Method: Principal Component Analysis. Rotation Method:
Varimax with Kaiser Normalization. Rotation converged in 6 iterations.

Component
Measurement items of shopping value
HV UV
0.85
I can relieve stress and feel relaxed when I shop online.
8
0.85
I enjoy passing the time by shopping online.
1
0.84
I look around at items on the Internet just for fun.
5
0.77
I feel like an escape when I shopping on this website.
9
I go online shopping because I can quick access to large 0.82
volumes of product and service information. 6
I go online shopping because online shopping saves time 0.81
compared to in-person shopping. 2
I go online shopping because I can have a more comfortable 0.76
and convenient shopping environment 6
I go online shopping because I can get the same quality 0.67
products at a lower price. 6
Note: Extraction Method: Principal Component Analysis. Rotation Method:
Varimax with Kaiser Normalization. Rotation converged in 3 iterations.

37
 Component
Measurement items of behavior
PI TR SAT
I will visit this web site when I want to buy certain items in 0.85
near future. 1
I will buy an item from this website if I find something that I 0.84
like. 4
0.74
I intend to acquire product information from this website.
3
I intend to purchase more goods from this website in the 0.71
next few years. 7
0.87
Most of what this web site says about its products is true.
1
If this web site makes a claim or promise about its product, 0.81
it is probably true. 8
0.69
This website is honest.
1
This website is trustworthy. 0.64
0.77
The online shopping website makes me feel very satisfied.
9
Using the online shopping website makes me feel very 0.74
pleased. 3
Using the online shopping website makes me feel very 0.66
content. 2
Note: Extraction Method: Principal Component Analysis. Rotation Method:
Varimax with Kaiser Normalization. Rotation converged in 4 iterations.

38