Chapter 2 Literature Review
2.3 Hedonic Value
Hedonic value, defined as consumption behaviors that relate to fantasy, happiness, sensuality, and
enjoyment (Hirschman and Holbrook, 1982), is an overall assessment of experiential benefits.
Compared with conventional utilitarian shopping values, the merit of hedonic value is experiential
and emotional. The reason why hedonic consumers do shopping is not for physical objective but for
the shopping process instead. Research concerning hedonic motivation becomes more popular lies in
two reasons. One is the obvious value that appeals to consumers to patronage the website, the other
is the fact that hedonic value is the extension of utilitarian value and these two values seem to
become crucial factors in keeping competitive advantage (Parsons, 2002).
Numerous studies used to adopt hedonic value dimensions to discuss in-store shopping.
Nevertheless, there are more and more research using hedonic value dimensions to explore online
shopping. Except for the freedom to search, hedonic value is also an important element. Mathwick et
al. (2001) discusses the experiential value of online shopping, and enjoyment and aesthetics should
be viewed as hedonic value. Kim and Shim (2002) propose that consumers go on line is not only for
information and products, but also for emotional satisfaction. Hedonic online shoppers are
accustomed to active pursuit while online. They often browse website, search for new items and
12
download updated information, actually they are bathed in the process of enjoyment. In addition,
hedonic shoppers were attracted to online environment because the virtual community offers social
relationship (Wolfinbarger and Gilly, 2001).
According to previous studies, it appears that hedonic values play important roles in online
shopping (Pui et al., 2007). Childers et al. (2001) indicated that hedonic orientations are strong
predictors of attitudes toward online shopping. Hedonic shopping values were positively related to
attitude toward online shopping and intentions to shop online (Menon and Kahn, 2002). In view of
the above, it is plausible to expect a positive relationship between hedonic values and attitude toward
online shopping. It is thus hypothesized that:
H14: Adventure is positively related to attitude toward online shopping.
H15: Sociality is positively related to attitude toward online shopping.
2.4 Main effects of gender on online shopping values
This study categorizes utilitarian value into convenience, availability of information and lack of
sociality, they are detailed as follows.
13
2.4.1 Convenience
Convenience is defined as time savings and effort savings, including physical and mental effort.
Convenience is a crucial attribute for consumers when shopping online.
Shopping online makes it easy for consumers to locate merchants, find items, and procure offerings
(Balasubramanian, 1997). Wolfinbarger and Gilly (2001) mentioned that internet shopping provides
a more comfortable and convenient shopping environment. Consumers do not have to leave their
home and they can also browse for items by category or online store. Schaffer (2000) argued that a
convenient internet shopping provides a short response time and minimizes customer effort.
Swaminathan et al. (1999) reported that male internet buyers were more convenience oriented
and less motivated by social interaction than women Internet buyers. Alreck and Settle (2002)
indicated that women have more positive attitudes toward shopping, while men prefer shopping via
internet (Alreck and Settle, 2002). Hence,
H21: Men have higher scores on convenience than women.
2.4.2 Information availability
Bakos (1997) postulated that the internet includes abundant public information resources that can
be easily collected. For online shoppers, internet is the most efficient means to get related
14
information. The internet as a medium facilitates searching both product specifications and price
information. Price is an important reference and adolescent consumers often compare price between
multiple websites.
Women tend to be more sensitive to related information online than men when making
judgments (Meyers-Levy and Sternthal, 1991), bringing out subsequent purchase attitudes and
intentions presented by men and women to differ. In other words, females make greater use of cues
than males. Cleveland et al. (2003) find that when executing consumption women seek more
information than men. Hence,
H22: Women have higher scores on information availability than men.
2.4.3 Lack of sociality
Wolfinbarger and Gilly (2001) indicate that online shopping enables people to execute a
transaction without contacting others, and online buyers have more freedom and control over the
transaction. One advantage of online shopping is that online buyers can decide to buy or not and the
transaction is under their control. In addition, while shopping online, it turns out that people can
avoid social interaction and crowded environment.
15
Since men and women differ significantly in conventional buying motivations, we seem to
assume that this would also be the case on the internet purchasing. Swaminathan et al. (1999)
reported that male internet buyers were less motivated by social interaction than women Internet
buyers. Compared with men, women tend to enjoy shopping (Alreck and Settle, 2002), and they can
have more social interactions in the process of consumption. Computer-mediated shopping does not
offer women much social contact. Hence,
H23: Men have higher scores on lack of sociality than women.
Based on Pui et al. (2007), the hedonic values in this study comprise adventure and sociality, they are
detailed as follows.
2.4.4 Adventure
Adventure refers to the fact that shopping can bring stimulation and excitement, meaning that
consumers can run across novelty and interesting affairs in the process of fantastic shopping
(Westbrook and Black, 1985). Experienced consumers are inclined to view the shopping experience
as thrills, excitement and amazement. Babin et al. (1994) regard adventurous aspect of shopping as
an element that may produce hedonic shopping value. Sherry (1990) addresses that in the shopping
process shoppers pay more attention to sensual satisfaction rather than the product itself.
16
Women stress emotional and psychological involvement in the whole shopping and buying
process, while men emphasize efficiency and convenience in obtaining buying outcomes (Dittmar et
al., 2004). In other words, the added value attached to shopping process may play a much more
prominent role for female consumers, whereas male consumers’ primary concerns are to get the
product only, shopping process may function as nothing meaningful for men. Hence,
H24: Women have higher scores on adventure than men.
2.4.5 Sociality
Sociality, grounded in McGuire’s (1974) collection of affiliation theories of human motivation,
suggesting that people put emphasis on cohesiveness, affiliation and affection in interpersonal
relationships. Tauber (1972) indicated that shoppers are fond of affiliating with reference groups and
interacting with those who have similar interests. Westbrook and Black (1985) regarded affiliation as
a shopping value, and Reynolds and Beatty (1999) stress the importance of social motivations for
shopping. Wolfinbarger and Gilly (2001) propose that virtual community is a new platform of
sociality, meaning that Internet shopper can share updated information and related shopping
experiences with one another and compared with traditional social benefits from friends, virtual
community furnishes shoppers with fresh pleasure.
17
Alexander (1947) mentioned that the experimental psychologists have developed very
convincing evidence that women are more prone to social contacts than men, and reasonably
convincing evidence that they have more aptitude for maintaining such contacts. Based on the results
of carefully conducted aptitude and interest tests, the gender differences seem to be notable.
Dittmar et al. (2004) address that women, compared with men, have a strong desire for emotional
and social gratification in the Internet buying environment. Hence,
H25: Women have higher scores on sociality than men.
2.5 Attitude toward online shopping
Attitude towards a behavior refers to ―the degree to which a person has favorable or unfavorable
evaluation of the behavior of the question‖ (Grandom and Mykytyn, 2004). Attitude can be viewed
as the bridge between consumers’ background characteristics and the consumption that fulfills their
needs (Armstrong and Kotler, 2000; Shwu-Ing, 2003). A person’s shopping choices are influenced
by attitude (Haque et al., 2006).
Attitude toward online shopping can be defined as a person’s positive or negative feelings about
conducting the purchasing behavior on the internet (Chiu et al., 2005; Schlosser, 2003). The consumers’ attitude towards online shopping plays an important role in determining e-shopping
18
behaviors (Shwu-Ing, 2003). Because attitude is difficult to change; therefore, to understand consumers’ attitudes toward online shopping enables marketing managers to predict the online
shoppers’ behaviors and satisfy their needs (Armstrong and Kotler, 2000).
2.6 Gender as a Moderator
Numerous studies related to consumer behavior indicated that males and females differ in their
processing of information (Holbrook, 1986; Palmer and Bejou, 1995). In other words, males and
females differ in the response to alternative consuming stimuli (Meyers-Levy, 1989).The differences
between males and females suggest the potential moderating role of gender in the influence of
shopping values (convenience, availability of information, lack of sociality, adventure, and sociality)
on attitude toward Internet purchasing, because online shopping causes different stimuli than those of
physical store. Therefore, males and females would vary in making judgments when facing relevant
information online (Meyers-Levy and Sternthal, 1991), and this may result in gender differences in
attitudes toward Internet purchasing. The following hypotheses are rooted in this gendered analysis.
H26: The relationship between online shopping value and attitude toward online shopping is
moderated by gender.
19
2.7 Gender research
The discrepancy between men and women really does exist, including physical and mental differences. The differences drew marketing researchers’ interest that brought out related studies,
such as gender differences in decision-making styles (Vincent and Walsh, 2004), attitudes toward
Internet, catalog, and store shopping (Alreck and Settle, 2002), attitudes toward Internet and store
shopping (Dholakia and Uusitalo, 2002), online and store buying motivations (Dittmar et al., 2004),
perceived risk of buying online (Garbarino and Strahilevitz, 2004), and Internet shopping behavior
(Chang and Samuel, 2004). Moreover, Jackson et al. (2001) noted that though young women and
men use the Internet equally often, they use it differently, and this may influence the motivations of
buying online. This steam of research may be a fertile area in marketing.
2.8 Summary
This chapter developed a conceptual framework based on technology acceptance model. The
literature review provided necessary and sufficient statements to support the hypotheses which
specify the relationships between the constructs: convenience and attitude toward online shopping,
information availability and attitude toward online shopping, lack of sociality and attitude toward
online shopping, adventure and attitude toward online shopping, and sociality and attitude toward
online shopping. In addition, gender has main effect on convenience, information availability, lack of
sociality, adventure and sociality, and the relationship between online shopping value and attitude
20
toward online shopping is moderated by gender. The following chapter illustrates the methodology
which will be used to examine the model.
21
Chapter 3. Methodology
This chapter presents the methodology used to examine the conceptual model. The following
addresses (1) the instrument design, (2) sampling, and (3) pretest and pilot study used in this study.
The details are presented as follows.
3.1 Instrument Design
The scales in this study are derived from previous studies, and are modified according to the
special conditions of Internet shopping. The scales are presented in the Appendix A. The
questionnaire is composed of three parts, the first part included 15 items to measure the utilitarian
and hedonic values of Internet shopping. The second part included four items measuring the Internet shoppers’ attitude toward online shopping. The last part collected the demographic data of the
subject. Detailed definitions of the dimensions are provided in the following text.
Utilitarian values
The items of convenience are derived from the questionnaire of Eastlick and Feinberg (1999) and
Pui et al. (2007). Information availability was assessed using a 3-item version of the scale developed
by Wolfinbarger and Gilly (2001) and Pui et al. (2007), while the items of lack of sociality are
adopted from the scale developed by Pui et al., (2007). Respondents were scored on a scale ranging
from 1 = total disagreement to 7 = total agreement. The internal consistency of each shopping value
22
met the standard of Cronbach’s α in excess of .7 with sufficient reliability (Nunnally and Bernstein,
1994).
Hedonic Values
The items of adventure value were adopted from the questionnaire of Arnold and Reynolds (2003),
whereas the items of sociality value come from the scale developed by Arnold and Reynolds (2003)
and Pui et al. (2007). Item responses were on a 7-point Likert-type scale, ranging from 1 = total
disagreement to 7 = total agreement. The Cronbach’s alpha coefficient for each shopping value
exceeded .7, indicating that the scale had adequate reliability (Nunnally and Bernstein, 1994).
Attitude toward Online Shopping
Attitude refers to a consumer's overall evaluation of online shopping as a way of shopping, which
can be positive / favorable or negative / unfavorable. The items used to measure the attitude toward
online shopping were modified from the scale developed by Taylor and Todd (1995). Attitude
toward online shopping was measured by four items. Respondents respond to the items using a
seven-scale Likert-type scale. The alpha coefficient of this variable was 0.88 with sufficient reliability
(Nunnally and Bernstein, 1994).
23
3.2 Sampling
The data came from Taiwanese universities. Respondents were approached in and around a
university community located in a major metropolitan area during October of 2009. Sampling is
conducted via convenient sampling. Study subjects were undergraduate students in Taiwan who have
shopped online. Notably, university students play an important role in online shopping and represent
a long-term potential market (Bruin and Lawrence, 2000). A self-administered questionnaire was
randomly distributed to 600 students attending the chosen school. Of the 308 questionnaires returned,
12 were incomplete. The remaining 296 were valid for analysis, representing a response rate of
49.33%. Among the 296 subjects, 168 (56.7%) were males and 128 (43.3%) were females.
24
Table 3.1 Characteristics of sample.
Characteristics n %
Gender
Male 168 56.7
Female 128 43.3
Age
19 68 22.9
20 82 27.7
21 86 29.1
22 60 20.3
Note: N = 296.
3.3 Pretest and Pilot Study
Pretest was adopted before conducting formal survey, confirming that the questionnaire had no
semantic problem. Two Ph.D. candidates majoring in marketing serve as the subjects of the pretests.
The pretest was performed in an open-end format that the subjects could propose any questions about
the items. During the process, the subjects suggested that the phrasing of certain items could be
revised, which included utilitarian value, hedonic value, and attitude toward online shopping.
According to several suggestions from the pretest subjects, we slightly revised wording of the items.
25
After the pretest, we conducted pilot test by distributing 60 questionnaires to some undergraduate
students having online shopping experience. A total of 30 questionnaires were received for reliability
analysis. This study examined Cronbach’s alpha of all constructs, and we eliminated some items that
the Cronbach’s alpha are below 0.7. The alpha coefficients of remaining items were above 0.7.
Finally, nineteen items remained in the questionnaire were used for later analysis.
3.4 Summary
This chapter provided information for instrument design and data collection. The measurement
scales for the constructs were adopted from previous studies. The following chapter will illustrate
statistical technique, structural model, fit indices of model fit and the result of the hypotheses testing.
26
Chapter 4. Analysis and Results
This chapter presents the results from the data analysis by using LISREL 8.51computer software.
The reliability and validity of the constructs are presented, followed by the discussion of model fit.
Finally, the hypotheses and the model are tested and the results are elaborated.
4.1 Analytical strategy
This study used structural equation modeling (SEM) to explore the relationships among the
variables. SEM enables researchers to definitely test research hypotheses concerning the
relationships among research constructs. This study used LISREL 8.51 to assess the fit of the
measurement and structural models, and calculated the values of incremental fit index (IFI) (Bollen,
1989), comparative fit index (CFI) (Bentler, 1990), normed fit index (NFI) (Bentler and Bonett,
1980), and root mean square error of approximation (RMSEA) (Steiger, 1990). If the values of IFI,
CFI and NFI exceed the cut-off value of 0.9, and the value of RMSEA is below the cut-off values
0.08, then the model is said to be acceptable (Hu and Bentler, 1999). In addition to the fit indices,
this study uses the parameter estimates in the structural model to test the hypotheses.
4.2 Reliability of Measures
To assess the internal consistency of the constructs, the Cronbach alpha values were examined
and were shown on Table 4.1. The result of six constructs ranged from 0.86 to 0.91, meeting the
27
lower standard of 0.7. In addition, the composite reliability (CR) shows the internal consistency of
the indicators assessing a given factor and is calculated by the formula suggested by Fornell and
Larcker (1981). If the value is higher than 0.7, we can conclude that the indicator is acceptable for a
composite reliability (Fornell and Larcker, 1981). As seen in Table 4.1, the CR scores of all
constructs exceeded the acceptable levels.
Table 4.1. Reliability of Measurement Scales
Construct Cronbach’s
Alpha
CR
Convenience 0.91 0.92
Information availability 0.89 0.90
Lack of sociality 0.88 0.90
Adventure 0.86 0.87
Sociality 0.86 0.88
Attitude toward online shopping 0.88 0.89
4.3 Validity of Measures
Before testing the hypotheses, this study first conducted average variances extracted (AVE) to
ensure that respondents can explicitly distinguish the variables in this study. The average variances
28
extracted (AVE) represents the amount of variance captured by the construct’s measures relative to
measurement error and the correlations among the latent variables. The analysis indicated that all
items had factor loadings higher than 0.7 (see Table 4.2). As seen in Table 4.2, all of the AVE values
were larger than 0.5, revealing good convergent and discriminant validity (Fornell and Larcker,
1981).
Table 4.2. Validity of Measurement Scales
Constructs Loading AVE
Convenience 0.76
I can buy things whenever I want. 0.83
I can buy things at home. 0.76
Online shopping is convenient to me. 0.80
Availability of Information 0.68
I can get information easily online. 0.86
Internet provides a lot of information. 0.79
Information via internet is the newest. 0.83
Lack of Sociality 0.75
I can avoid embarrassment when I buy things online. 0.87
Online makes me free from salesman. 0.81
29
Online makes me free from social interaction. 0.84
Adventure 0.64
Online shopping is an adventure. 0.90
I find shopping stimulating. 0.81
Online shopping makes me feel like I am in my own universe. 0.85
Sociality 0.67
I can exchange information with friends online. 0.78
I can develop friendships with other internet shoppers. 0.83
I can extend personal relationship online. 0.81
I can exchange information with friends online. 0.85
Attitude toward online shopping 0.72
Using the Internet to buy things is a good idea. 0.88
I like the idea of busing what I need via the Internet. 0.81
Using the Internet to buy things is a wise idea. 0.86
Using the Internet to buy things would be pleasant. 0.82
Overall, all the constructs show very good reliability and validity. Since we confirmed the
reliability and validity of constructs, we then test our hypotheses by estimating the full model. A
structural model was used to test the previously presented hypotheses. SEM results for testing the
30
model depicted in Figure 4.1 revealed that the fit statistics of model were within the recommended
range (chi-square value= 310.27, d.f.=120 (p < 0.001), RMSEA=0.076; CFI =0.93; IFI=0.93;
NFI=0.90). All of these fit indices are within the acceptable limit, suggesting that the overall
structural model provides a good fit with the data. Owing to an excellent fit between the structural
model and the data, this study then tested the hypotheses according to its parameter estimates.
Figure 4.1. Original model.
31
4.4 Hypothesis Test
Hypothesis 11-15 depicted that online shopping values are positively related to attitude toward
Internet shopping. As observed, Table 4.3 represents summaries of the hypothesis tests. The
structural links to attitude toward Internet shopping from convenience (H11), availability of
information (H12), lack of sociality (H13), adventure (H14), and sociality (H15) are completely
supported. Therefore, the model is supported as five online shopping values positively influence
attitude toward online shopping, indicating that hypothesis 11-15 were fully supported. In addition, this study divided the sample respectively by men and women, further investigating across men’s
and women’s groups for latent means testing and subgroup analysis.
Table 4.3 Results of hypotheses and model statistics.
Hypotheses Path
32
shopping
H15: Sociality Attitude toward online
shopping
0.18 3.43*** Supported
***p < 0.01.
4.5 Latent Means Difference
Based on the above measurement model, latent means testing is conducted (Byrne, 2001). Table
4.4 represents the results of latent means testing. Given that the women group functions as the reference group, the members’ factor means are hence fixed to zero to examine the latent means
difference between the two subgroups. As seen in Table 4.4, the significantly positive estimate of the
latent means difference across the subgroups for construct H21 and H23 indicates that the scores on
convenience, lack of sociality are significantly higher for men than for women. On the other hand,
the significantly negative estimate of the latent means difference across the subgroups for construct
H22, H24 and H25 indicates that the scores on availability of information, adventure and sociality
are significantly higher for women than for men.