Journal of Business & Industrial Marketing

How does word of mouth Affect customer satisfaction?

Wenhua Shi Lingshu Tang Xiaohang Zhang Yu Gao Yameng Zhu
Article information:
To cite this document:
Wenhua Shi Lingshu Tang Xiaohang Zhang Yu Gao Yameng Zhu , (2016),"How does word of mouth Affect customer
satisfaction?", Journal of Business & Industrial Marketing, Vol. 31 Iss 3 pp. -
Permanent link to this document:
http://dx.doi.org/10.1108/JBIM-07-2014-0139
Downloaded on: 11 March 2016, At: 08:07 (PT)
References: this document contains references to 0 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 30 times since 2016*
Users who downloaded this article also downloaded:
Ravi Pappu, Pascale G. Quester, (2016),"How does brand innovativeness affect brand loyalty?", European Journal of
Marketing, Vol. 50 Iss 1/2 pp. 2-28 http://dx.doi.org/10.1108/EJM-01-2014-0020
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

Peter A. Voyer, Chatura Ranaweera, (2015),"The impact of word of mouth on service purchase decisions: Examining risk
and the interaction of tie strength and involvement", Journal of Service Theory and Practice, Vol. 25 Iss 5 pp. 636-656 http://
dx.doi.org/10.1108/JSTP-04-2014-0070

Access to this document was granted through an Emerald subscription provided by emerald-srm:121184 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service
information about how to choose which publication to write for and submission guidelines are available for all. Please
visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of
more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online
products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication
Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.

*Related content and download information correct at time of download.

 HOW DOES WORD OF MOUTH AFFECT CUSTOMER

SATISFACTION?

Introduction

Imagine going to Hai Di Lao Hot Potto enjoy a dinner. Because friends have

praised its serviceand conveyed positive word of mouth (WOM), you look forward to

enjoying the services of Hai Di Lao Hot Pot with relatively high expectations.
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

Unfortunately, your waiter is in a bad mood and provides you with poorer service than

you expected. Are you dissatisfied? Or will you insist that the perceived quality is

high and that this service failure is the result of accidental factors, so you will

eventually give a high service evaluation? To attempt to solve this problem, we

explore the relation between WOM and customer satisfaction.

There are numerous studies on the relationship between WOM and customer

satisfaction. In the majority of the literature, the degree of satisfaction is an

explanatory variable of WOM; specifically, the majority of research on this topic

focuses on how the consumer’s satisfaction affects the spread of WOM. When

consumers receive services that exceed their expectations, consumers are satisfied and

are willing to spread positive WOM regarding the service provider to others.

Conversely, the less satisfied the consumer is, the more likely he is to generate

negative WOM(Anderson, 1998; Ranaweera & Prabhu, 2003; Luo& Homburg et al.,

2007).

Some of the literature considers WOM as an explanatory variable, including how

 WOM affects the expectations of customers (e.g.,Parasuraman, Zeithaml & Berry,

1985; Chen, Liu, & Zhang,2012), how WOM influences perceived quality

(e.g.,Hartline & Jones, 1996; Wangenheim,2010), and the influence of WOM

information on consumers’ post-purchase satisfaction (Ren,2010).

Based on existing theory, satisfaction with Hai Di Lao Hot Pot may develop in

one of two ways: 1) the positive WOM from friends may raise expectations of Hai Di

Lao’s service (Zeithaml, Berry & Parasuraman, 1993); however, because of the bad
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

waiter, the consumer’s expectations are higher than the perceived quality, resulting in

a lower degree of satisfaction; or2) positive WOM may raise the consumer’s

expectation of the service at Hai Di Lao Hot Pot. Simultaneously, as a result of the

halo effect, the customer may insist that this bad experience is just an anomaly. When

the consumer has high expectations, the perceived quality is higher. Furthermore, if

only one person is unsatisfied, that person will be the subject of interpersonal/group

pressure to ultimately ensure the dissenter’s satisfaction. To explore real results, we

focus on how WOM (positive and negative) received prior to a purchase affects

customer satisfaction. Theoretically, understanding this relation can enrich the

corresponding study of the relations between WOM and satisfaction. These results

can also provide management with certain practical guidelines from the perspective of

service, such as the risks of spreading positive WOM.

Although the focus of the study of WOM has changed, according to Ren (2010),

the research has developed in two areas: first, among different consumption objects,

such as buying customers, potential customers and related groups; second, the
 different types of WOM, including positive WOM, negative WOM, Internet WOM

and mixed WOM. By exploring the different types of WOM, this research attempts to

explore how WOM affects the satisfaction of potential consumers and identifies the

inner mechanism of the effect of WOM on satisfaction. In this article, we define

WOM as a binary-type variable. The flourishing development of electronic commerce

has allowed the development of WOM information that is based on network platform

evaluations of a product or service. Electronic evaluations influence the attitudes and

Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

behaviors of customers in ways different from traditional offline WOM (e.g., Bickart

&Schindler, 2001; Henning, 2003; Pavlou, 2006; Chevalier, 2006). Thus, this study

will explore whether a different source of WOM changes the extent to which WOM

influences the actions of consumers (online WOM compared with offline WOM).

Consumers focus on different aspects of different types of products. Bei (2004) noted

that when consumers make a decision to buy experience goods, WOM is more

important than when they buy search products. With search goods, the consumer pays

more attention to actual perceived quality during use than he does with experience

goods. Hence, this study will also include product type and an analysis of the role

product type plays in the relation between WOM and customer

satisfaction (experience goods compared with search goods).

This article is organized as follows. After developing our conceptual framework

and the corresponding hypotheses, we describe the methodology and the results of the

empirical studies. Next, we explain the results and discuss research-related and

managerial implications.
 Literature review

Our aim in this study is to discuss how pre-purchase WOM affects the

satisfaction of customers. WOM, a binary variable, can be divided into positive WOM

and negative WOM. Satisfaction refers to the attitude of customers toward a product

or service after receiving purely positive or negative WOM; the customer then has a

practical experience .Because WOM affects consumers’ expectations and perceived

quality (Zeithaml, Berry & Parasuraman, 1993; Buttle, 1998) and because
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

expectations and perceived quality are two important antecedent factors of customer

satisfaction (Oliver, 1981), we have reason to believe there is a relationship between

WOM and satisfaction. More importantly, we consider the source of WOM and

product type to be moderating variables in influencing the WOM/satisfaction relation.

In our literature review, we first review the relationship between WOM and

customer satisfaction, including the research on the influence of customer satisfaction

on WOM and the basic theory of the two possible paths by which WOM can influence

customers’ satisfaction. Then, we present a literature summary to support the source

of WOM and product type as moderating variables.

The Relationship between Word of Mouth and Satisfaction

Satisfaction-Word-of-Mouth

WOM is informal communication among consumers regarding the supplier

and/or its product/service characteristics. WOM can be positive or negative (Tax,
Chandrashekaran & Christiansen, 1993) and is received by potential customers before
they experience a product or service. Research on the effect of satisfaction on WOM
is abundant. Based on previous literature, six dependent variables—customer
 satisfaction, customer loyalty, commitment, trust, quality, and perceived
value—influence the generation of WOM. Based on traditional theory, customer
satisfaction occupies a decisive position in customers’ long-term buying decision
behavior; the more satisfied a customer is, the more WOM information she is likely to
produce. When customers consume or experience a product or service, they will
compare their expectations with the product or service and, according to their own
experience, will evaluate the actual situation of the product or service. When the
actual situation cannot meet expectations, the customer will be dissatisfied; on the
contrary, when the actual condition meets or exceeds the customer’s expectations, the
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

customer’s satisfaction with the product or service will be high. Customer satisfaction
has an important influence on the spread of WOM.
Swan and Oliver(1989) believed that satisfied customers would produce more
WOM than unsatisfied customers. Many studies have confirmed the positive
correlation between satisfaction and WOM. As Ranaweera and Prabhu (2003)
suggested, satisfaction and positive WOM have a strong relationship; unsatisfied
customers are more likely to produce negative WOM, and satisfied customers are
more likely to spread positive WOM. In addition to the linear relationship, Anderson
(1998) confirmed that there is a u-shaped relationship between customer satisfaction
and WOM. Specifically, when the degree of satisfaction is higher or lower, consumers’
intention to spread WOM is higher, and when customers’ satisfaction or
dissatisfaction is average, their intention to spread WOM is lower.

As seen from the existing research results, in the research on WOM and

satisfaction, satisfaction is always an explanatory variable. The research topic of how

customer satisfaction affects customers’ WOM information sources is mature.

Word of Mouth Expectations

In the model of the satisfaction and service quality gap (5 Gap Model) created by

Parasuraman, Zeithaml and Berry (1985), WOM (oral communication) is one of the
 factors that influence the expectations of the consumer at the customer level.

Thestudies on WOM and consumer expectations have been relatively in depth.

Expectation is the "level at which the customer wants the product to perform" (Swan

& Trawick, 1980) and reflects how well the focal words of expressions fulfill the

innate needs, wants, or desires of consumers. Lynn, Stock, Zinsner and Woodside

(1987) confirmed that in the purchase of industrial goods, WOM affects the

expectations of buyers at the search stage and influences the attitudes of the buyers at
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

the select evaluation phase. Based on the theory of Wilson and Peterson (1990), if the

consumer’s attitude toward a product or service is initially positive, then the

consumer’s attitude will be strengthened by positive WOM regarding the product or

service, which indicates that expectations of the product or service will grow. If

consumers have a negative attitude toward the product or service, negative WOM will

strengthen this type of negative attitude and reduce expectations. Similar conclusions

have been confirmed in the study by Zeithaml and other scholars. Positive WOM can

promote customers’ expectations, and negative WOM can reduce customers’

expectations of a product or service (Zeithaml, Berry & Parasuraman,

1993).Third-party product reviews play significant roles because they lead investors

to update their expectations of a new product’s sales potential (Yong Liu, & Jurui

Zhang, 2012).

Word of Mouth Perceived Quality

From a psychological point of view, the halo effect is a cognitive bias in which

one’s judgments of a person’s character can be influenced by one’s overall impression

 of the person. If the cognitive object is considered "good" (positive WOM), the object

will be assigned all good qualities; if the cognitive object is considered "bad"

(negative WOM), the object will have a "bad" aura, and all of its qualities will be

considered bad (Edward Lee, Thorndike, 1920s). The “good qualities” or “bad

qualities” can also be called perceived qualities. Perceived qualities are customers’

actual feelings after using a product/service at a certain time, including the degree to

which the product/service meets individuals’ specific demand and the level at which
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

consumers perceive a product’s reliability and quality.

Several studies have supported the halo effect (Nisbett & Wilson, 1977; Wirtz &

Bateson, 1995). In addition, studies have shown that WOM and perceived quality

have important connections. Existing studies focus on two basic aspects. First, does

WOM affect perceived quality? Second, does WOM influence consumers’ perceived

quality positively or negatively? Hartline and Jones (1996) note the correlation

between WOM and perceived quality: positive perceived quality will make the

customer disseminate more positive WOM. Francis A. Buttle (1998) proposed that the

public praise of expectations, perceived quality, and customers’ attitudes and behavior

are influenced by WOM, and the influence of WOM on perceived quality is positive.

Based on cross-cultural studies, Schumann and Wangenheim (2010) suggested that

WOM positively influences customers’ perceived quality, and this effect is influenced

by cultural factors. In different cultural environments, the size of the effect is

different.

Through the above review, we see that WOM is an important factor that
 positively affects the quality of customers’ expectations and perception. Accepting

positive WOM can improve the quality of consumers’ expectations and perceptions,

whereas accepting negative WOM can reduce consumers’ expectations and

perceptions. Thus, we propose the following.

Hypothesis 1-1: The influence of positive WOM on customers’ expectations is

significantly different from that of negative WOM.

Hypothesis 1-2: The influence of positive WOM on customers’ perceived

Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

quality is significantly different from that of negative WOM.

As supported by the American Customer Satisfaction Index Model (ACSI),

customers’ expectations should be positively related to perceived quality and

customer satisfaction, and perceived quality positively affects satisfaction (Fornell,

Johnson, &Anderson, 1996). In addition, according to the Expectancy Disparity

Model (Oliver, 1980), expectations and perceived quality are intermediary variables

of WOM and satisfaction (Parasuraman, Zeithaml, &Berry, 1985; Fornell, Johnson,

&Anderson, 1996).

Thus, WOM affects satisfaction in two ways: 1) WOM positively influences

consumers’ expectations and negatively affects satisfaction; 2) by positively

influencing consumers’ actual perceived quality, WOM ultimately affects satisfaction

negatively. The two paths point in opposite directions. In addition, satisfaction is a

function of consumers’ actual perceived quality and expectations. Thus, we assume

the following.

Hypothesis 1-3: WOM can affect satisfaction.

 Hypothesis1-4: The influence of positive WOM on customer satisfaction is

significantly different from that of negative WOM.

The Moderating Effect of the Source of WOM

Depending on the source, WOM can be divided into offline WOM and online
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

WOM. Offline WOM is also called traditional WOM. According to Dong and Liu

(2012),it is a two-way, non-commercial oral communication that occurs between

consumers who have a strong relationship. The content is focused on a product (or

service) or is in some manner related to consumer behavior. Online WOM, also

known as electronic WOM, is spread by potential, actual, or former customers via the

Internet. This type of communication describes a product or enterprise with a positive

or negative statement and is observed by many people (Thorsten; Gwinner; & Walsh,

2004). Regarding trust, scholars have different viewpoints regarding the persuasive

power of online and offline WOM. Some scholars think that offline WOM occurs

between people who have a stronger relationship and can trust one another; thus, their

advice or opinions can affect consumers’ purchasing decisions to a greater extent.

Conversely, some characteristics of online WOM, such as communication between

people who have weak relationships, sender anonymity, and the spread of Internet

rumors, can affect the effectiveness of online WOM (Bansal & Voyer, 2000; Brown &

Reingen, 1987). Other scholars believe that in the virtual world, the anonymity of

those who provide opinions ensures their privacy, encourages free speech and makes
 their WOM more reliable (Sun, 2006). Thus, different WOM sources produce

different levels of credibility of WOM. According to the theory of rational behavior

(Fishbein and Ajzen, 1975) and Beck (2007), the credibility of WOM affects

consumers’ attitudes and purchasing intentions. Moreover, according to the

prestige-of-communication effect, when the audience perceives that the source of

WOM information is highly authoritative and reliable, the audience will believe the

content of the information. The authority of the information spreader positively

Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

affects the audience’s attitude (Schramm, 2010). We believe that the more consumers

trust the origins of WOM, the more consumer expectations and perceived qualities are

affected. Thus, we propose that the source of WOM moderates the effect of the

received WOM on customers’ expectations and perceived qualities, influencing the

relationship between WOM and customer satisfaction.

Hypothesis2-1:The source of WOM affects the relation between WOM and

expectations.

Hypothesis2-2: The source of WOM affects the relation between WOM and

perceived quality.

Hypothesis2-3:The source of WOM affects the relation between WOM and

satisfaction.

The Moderating Effect of Product Type

Based on the theory of information search, the information economist Nelson

(1970) suggested that products are divided into two categories: experience goods and

search goods. This classification method has been adopted by some scholars in the
 study of WOM in relation to product categories (e.g.,Mitra K, McCain & Capella,

1999). Experience goods refer to a class of products that a consumer must purchase

and use before fully confirming the quality(Klein, 1998).Conversely, consumers can

obtain information on the features of search goods before buying or using them. This

information helps consumers determine the quality of products and their applicability

(Nelson, 1974). In research on WOM and product types, Yong (2005) noted that the

influence of WOM is different for different types of products. Experience goods, such
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

as movies and other entertainment products, are more likely to be affected by WOM.

Bei et al. (2004) observed that consumers tend to use WOM more frequently when

buying experience goods than when buying search goods. Clearly, different types of

products depend on WOM to different degrees. Jiet al. (2009) observed that

consumers tend to rely on WOM to render a judgment regarding experience goods

and services, whereas the evaluation of search goods relies more on understanding the

properties of the goods (Li, Zhang, 2013). Thus, we propose that in comparison with

search products, expectations and the perceived quality of experience goods rely more

on WOM, which is also likely to modify the relation between WOM and customers’

expectations and perceived quality.

Hypothesis3-1: Product type affects the relation between WOM and expectations.

Hypothesis3-2: Product type affects the relation between WOM and perceived

quality.

Hypothesis3-3: Product type affects the relation between WOM and satisfaction.
 Study Design

The data collection method for the experiment can solve the problem of time

intervals and is appropriate for isolating causality, manipulating independent variables,

and measuring or observing subsequent results (Perdue, &Summers, 1986). To

address our hypotheses, we manipulated the content of WOM(positively or negatively)

in a field-based experimental setting and measured expectations, perceived quality

and customer satisfaction. Based on two sets of WOM information, we divided the
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

experiment into offline and online WOM scenarios. Products included experience

goods and search goods.

We conducted our experiment with 160 college students in classrooms. We

measured the students’ expectations, perceived quality of products and satisfaction.

We manipulated three factors among the subjects. The first factor concerned the

content of WOM, including positive and negative WOM, for which the number and

the strength were consistent. According to Sun (2009), despite the abundance of

Internet WOM, most consumers carefully read approximately 10 sources of

information before purchasing a product. Thus, we prepared 10 positive and 10

negative examples of WOM. To control the strength of the positive and negative

WOM, we created positive and negative WOM that correlated with one another, such

as “This product is very good” and the corresponding negative WOM, “This product

is very bad”. The second factor involved the source of WOM. Some WOM spreads

online. WOM information regarding experience goods is derived from the Internet, so

we showed WOM in the form of screen shots(Lai,2010). Some WOM spreads offline;
 WOM information regarding experience goods is derived from participants’ friends

and acquaintances. The third factor involves product type. For experience goods, we

used a type of drink, and for search goods, we used a mobile phone. We chose these

products because our research subjects were primarily college students who were

similar to one another; they easily accepted new things and explored new products. A

beverage is an example of a typical experience good and a daily necessity. The

majority of a beverage’s attributes can be understood only after purchasing and using
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

the product. A mobile phone is a typical example of a search good. College students’

spending power is not great; when they make a purchasing decision, WOM is an

important reference factor. Additionally, according to Li (2005), Chinese consumers

have the greatest dependence on WOM when buying mobile phones and air

conditioning systems.

To minimize the information conveyed by the first experimental groups to the

other groups of participants, we chose students from 8different colleges and organized

8 different experiments (positive compared with negative WOM, offline compared

with online, experience goods compared with search goods).Each group of activities

occurred in a closed classroom, and subjects maintained a certain distance from one

another to ensure that they could not communicate. During the entire experiment

process, we blocked the information regarding brands and selected newly issued

products that did not have a high degree of recognition among consumers. This

approach was employed primarily to ensure that the participants were influenced only

by the WOM we provided and were not subject to prior experience or brand factors.
 In addition, to ensure the validity of the data collection, we utilized a WOM

credibility test that adopted a 5-pointLikert scale (1 indicating “strongly disagree”, 5

indicating “strongly agree”). The credibility test included the credibility of the content

of the WOM(“I think the WOM I received was a real description of the product”)and

the source credibility (“I think the Internet friend/friends who spread the WOM are

trustworthy”). Questionnaires that included samples that scored below 2 points were

voided. We utilized the positive online WOM regarding experience goods to describe
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

our experiment process.

First, we collected ten positive pieces of information describing the good taste of the

beverage from three mainstream e-commerce sites (Amazon, Taobao, and Jingdong)

and displayed the information on a large screen. For offline word of mouth, we

invited ten student subjects (undercover) to taste the beverage at the scene of the

experiment. Next, the student subjects communicated WOM information regarding

the product to the other participants. Of course, the ten tasters were given the WOM

we had prepared, and the information content was completely consistent with the

online word of mouth. After reading the WOM, we had the participants fill in the

questionnaire regarding the credibility of the WOM content and source. The first two

questions were to ensure the validity of the data; scores of less than two points were

voided. After the participants completed the first questionnaire, we offered them the

beverage to taste. Next, we provided the second questionnaire for subjects to complete.

The second questionnaire was primarily used to measure the participants’ actual

perceived quality of and satisfaction with the drink. All questions on the
 questionnaires came from ACSI and used a 5-point Likert scale.

The final sample was structured as follows: 160students participated in this

experiment. We excluded cases in which respondents failed the manipulation check or

gave unreasonable answers. Specifically, we first excluded cases with many missing

values. Second, we rigorously analyzed the WOM credibility measurements to ensure

that the WOM really could influence the participants. The final sample consisted of

147 complete responses. The participants’ average age was 22.8 years, and the gender
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

balance was reasonable (females 49.6%, males 50.4%).Except for the offline search

product groups, the rest of the groups’ samples ranged from 20 to 24. Detailed sample

information can be found in table 1.

Measure

We used "completely agree/completely disagree" to measure expectations,

perceived quality and satisfaction. Each latent variable had three items, all from the

ACSI model-specific questionnaire.

The analysis of the 147 valid questionnaires showed that all Cronbach’s alpha

coefficients in this study were greater than 0.7, indicating acceptable reliability. To

assess measurement validity, confirmatory factor analyses were run for each reflective

measure individually using the software SPSS 19.0. Table 2 shows the reliability and

confirmatory factor analysis validity along with the measurement information. All

factors met this requirement. As shown in table2, the Cronbach’s alpha coefficients

were all higher than 0.7, and the standardized loadings of individual items were all

higher than or at the acceptable level of 0.60.The values of the average variance
 extracted were also satisfactory.

Results

First, we verified the presence of significant differences among expectations,

perceived quality and satisfaction by a multivariate analysis of varianceof the different

WOM content. Then, we used a structural equation model to explore the internal

mechanisms of the effect of WOM on satisfaction using the two types of WOM and
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

compared the path coefficient to test whether there was a significant difference.

Finally, we examined the moderating effect of the source of WOM and the product

type.

Multivariate Analysis of Variance

We tested H1-1, H1-2 ,H1-3 and H1-4 by analyzing between-group variation.

According to Hairet al. (2010), when three dependent variables are correlated to some

extent, using a multivariate analysis of variance (MANOVA) instead of three separate

analyses of variance (ANOVAs) is more reasonable, and a MANOVA has a higher

statistical power than separate ANOVAs for each of the dependent variables(Hair et

al., 2010). In our study, there are causal relations among customers’ expectations,

perceived quality and satisfaction. Thus, we performed a MANOVA with three

dependent variables: expectations, perceived quality and customer satisfaction. The

independent variable was WOM.

In this study, we measured the mean of the three items to represent the value of

each variable. The results of the MANOVA show a significant main effect of WOM
 on expectation (F(1,145)=115.59,p<.000) and customer satisfaction (F(1,

145)=8.651,p<.01). However, there is no main effect of WOM on perceived quality (F

(1, 145) =0.023, p>.05). Furthermore, expectations are higher when the WOM is

positive than when it is negative, and customer satisfaction is lower when WOM is

positive than when it is negative. Thus, H1-1 、H1-3and H1-4 are supported, and

H1-22 is rejected (table3).

Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

Influence Mechanism Analysis

We estimated the main effects of the structural equation model using the

software AMOS 17.0.

Figure 2 shows the structural model with the path coefficients of the positive

WOM. The goodness-of-fit indices of the proposed model were χ2/df=2.368,

CFI=0.920, GFI=0. 869, IFI=0.923, and RMSEA=0.094. Bentler (2000) notes that if

the CFI value reaches0.90 or above, GFI≥0.80 indicates that the model is acceptable

and that the evaluation of the model yields satisfying results.

Figure 3 shows the structural model with the path coefficients of the negative

WOM. The evaluation of the model also yields satisfying results. The goodness-of-fit

indices of the proposed model were χ2/df=2.318, CFI=0.929, GFI=0. 875, IFI=0.921,

and RMSEA=0.092.

In both the positive WOM and negative WOM groups, customers’ expectations

positively affected their perceived quality and negatively influenced their satisfaction.

Perceived quality positively influenced satisfaction.

 The Significance Test of the Path Coefficient

To explore whether there was a significant difference in the influence coefficient of

satisfaction between the different types of WOM, we conducted a path coefficient

difference significance test(Kei M. l, Tan B.C.Y. et al., 2000). The formula used to

calculate differences appears in Appendix A. For different types of WOM, the results

indicate that there was no significant difference in the expectation effect on

satisfaction(t=-0.46<t0.05 (145)=1.976)or the influence of perceived quality on

Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

satisfaction(t=0.76<t0.05 (145)=1.976).

The Test of the Moderating Effect

To test the moderating effect of the source of the WOM and product types, we used

a multivariate analysis of variance (MANOVA) because in this study, the type of

WOM is a binary variable that is separated into positive and negative WOM. In

addition to the independent variable, the moderating variables, the source of WOM

(online, offline),and the product type (experience goods, search goods) are all

binary-type variables. However, because the dependent variance (expectation,

perceived quality and customer satisfaction) is a continuous variable, we can use the

analysis of variance to analyze the interaction effect of the moderating variables and

the independent variables as long as the obvious interactions can be considered an

obvious moderating effect (Hillsdale, Erlbaum, 1983; M. Baron & A. Kenny, 1986).

We examined the regulating effect of the source of the WOM. First, we checked

whether there was a regulating effect of the source of WOM on the relation between

WOM and expectation and the relation between WOM and perceived quality. Then,
 we identified whether the relation between WOM and satisfaction changed because of

the different source of the WOM. The results indicate that the moderating effect of the

source of the WOM on adjusting the relation between WOM and the expectations of

customers is not significant (FEX (1,143)=0.48, p>0.05). The source of the WOM also

has no obvious moderating effect on the relation between WOM and customers’

perceived quality (FPE (1,143)=0.005, p>0.05); thus, the relation between WOM and

satisfaction is not modified by the WOM source (FSA (1,143)=0.006, p>0.05).More

Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

intuitively, in figure 3-1, we can see that the interaction effects are not significant;

thus, H2-1, H2-2and H2-3are refuted.

We also tested the regulating effect of product type. The results show that the

product type did not play a regulating role in the relation between WOM and

expectations (FEX (1,143)=2.08, p>0.05). However, product type had a regulating

effect on the relation between WOM and perceived quality (FPE(1,143)=5.686,

p<0.05). Thus, the relation between WOM and customer satisfaction was modified by

product type (FSA (1,143)=6.754, p<0.05), as shown in figure 3-2. The result is

favorable because H3-2 and H3-3 are supported, although H3-1is rejected.

Conclusions and Discussion

Conclusions

Although numerous prior studies have investigated the link between customer

satisfaction and WOM (Swan & Oliver, 1989;Anderson, 1998; Ranaweer & Prabhu,

2003; Luo & Homburg, 2007), investigations have neglected to acknowledge the

existence of the potential reverse effect of WOM on customer satisfaction. Utilizing

 an experiment that covers positive and negative WOM, offline and online WOM,

experience goods and search goods ,this article extends previous service research by

suggesting that WOM has a major influence on the satisfaction of customers. This

article also explores the internal influence mechanism of WOM’s effect on

satisfaction and confirms the moderating effect of product type.

Our results indicate that WOM can affect satisfaction; however, the path is from

WOM to expectations and then to satisfaction, not through perceived quality affecting
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

satisfaction. WOM positively affects expectations, and expectations negatively

influence satisfaction. Positive WOM can create higher expectations than negative

WOM, resulting in lower levels of satisfaction. Furthermore, the source of WOM

does not play a regulatory role in the relation between WOM and satisfaction. Product

type does play a regulatory role in the relation between WOM and perceived quality

and subsequently regulates the relation between WOM and satisfaction.

Discussion

The Mechanism of Word of Mouth Influence on Satisfaction

Variance analysis indicates that WOM significantly affects expectations. The

expectations of the positive WOM group (MP-E = 3.59) are significantly higher than

the expectations of the negative WOM group (MN-E = 2.23). These results are

consistent with the existing academic research (e.g., Wilson & Peterson, 1990;

Zeithaml, Berry & Parasuraman, 1993). When a consumer is deciding whether to

accept the influence of positive WOM or negative WOM, the perceived quality makes

no significant difference. This result is not consistent with the halo effect and existing
 research (Buttle, 1998; Wangenheim, 2010) for several reasons. First, the halo effect

refers to a type of psychological phenomenon in which the consciousness

indiscriminately extends an impression to other aspects while evaluating a

communication object. This affects the understanding of the essential characteristics

of an object and its evaluation. The halo effect develops during interpersonal

interaction (Xie, 2007). However, the experimental environment controls the

interaction among people so that the halo effect cannot influence the results. In
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

addition, in this study, we blocked brand names and selected new products for the

experiments that had not formed a large user base in the market, thus effectively

avoiding the effect of participants’ previous experience with these products.

Participants’ perceived quality of the experiment was immediately formed from this

single perceived quality after the experience with the product. The perceived quality

depended more on the degree of consistency of the product and the preferences of

customers than WOM. Thus, WOM had no marked effect on perceived quality.

Finally, by influencing expectations, WOM affected consumer satisfaction. The

satisfaction of the positive WOM group (MP-S=2.33) was lower than the satisfaction

of the negative WOM group (MN-S=2.75).

Specifically, to further explore the internal mechanism of the effect of WOM on

satisfaction, we compared the path coefficients derived from a structural equation

mode of expectation, perceived quality and satisfaction regarding positive and

negative WOM. The results of SEM show that regardless of whether the WOM is

positive or negative, expectations negatively influence consumer satisfaction (βP

 =-.259，βN =-.161), and perceived quality positively influences customer satisfaction

(γP=.47，γN=.354). To further ascertain whether there are significant differences in the

ways in which different types of WOM affect satisfaction, we identified a significant

difference in the path coefficient (Keil, Tan, 2000).The verification results show that

there is no significant difference in the path coefficients of the two groups. The results

show that for different types of WOM (positive WOM/negative WOM), the effect of

expectations on satisfaction is identical. In addition, there is no difference between

Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

positive WOM and negative WOM in the degree to which quality influences

satisfaction. Thus, there are no significant differences between positive WOM and

negative WOM regarding how expectations and perceived quality influence

satisfaction.

In conclusion, the path from WOM to expectation to satisfaction is remarkable,

and the effect of the path of WOM to perceived quality to satisfaction warrants further

investigation. The influence of WOM on perceived quality is not so obvious. However,

perceived quality has a strong influence on satisfaction, showing almost no difference

from the way in which expectation affects satisfaction. Eventually, the effect path

goes from WOM to expectations and then to satisfaction. In brief, positive WOM

promotes expectations and reduces customer satisfaction, and negative WOM

decreases expectations and improves customer satisfaction.

Moderating Effect

(The moderating effect of the source of WOM) The results of the analysis of the

moderating effect of the WOM source show that the relations between WOM and
 expectation, WOM and perceived quality, and WOM and satisfaction do not change

because of the source of the WOM (figure 4-1). In this study, the source of the WOM

is not a moderating variable. There has been much debate on the effect of online

WOM and offline WON on consumers. From the perspective of content, researchers

believe that for online WOM, the network environment is more conducive to

presenting true WOM without conformity pressure for consumers. Thus, online WOM

can have a greater effect on the consumer (Gem, 2002). However, some scholars
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

believe that offline WOM is more trustworthy from the perspective of WOM

communicators. Because offline WOM spreads among people connected by strong

relationships, it can more strongly affect the consumer (Johnson, 2000). Both views

are compelling. Our experimental results indicate that there are no significant

differences between online and offline WOM in terms of the effect on consumers’

expectations, perceived quality and satisfaction. Perhaps the different directions of the

two views modify the moderating effect. Furthermore, in this study, we provided only

the WOM and blocked any other relevant information regarding the experiment. Thus,

the expectations of the participants arose only from Internet WOM messages or WOM

messages from friends. Because both sources are strongly persuasive and credible, it

did not matter to the participants whether this WOM information came from the

Internet or from offline communication between friends; whether the WOM content

was positive or negative was the main factor in consumers’ expectations, a finding

that this study confirmed (MP-E=3.59, MN-E=2.23, F (1,145)=115.59, p<. 000).At the

same time, because the relationship between WOM and satisfaction is established by
 expectations and perceived quality, when the WOM source does not modify both the

relation between WOM and expectations and WOM and perceived quality, the source

will also not affect the relation between WOM and customer satisfaction. In this study,

therefore, the WOM source does not modify the effect.

(The moderating effect of product type)The results of the analysis regarding the

adjustment effect of product type suggest that product type significantly influences

the relation between WOM and perceived quality and between WOM and satisfaction.
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

However, there is no moderating effect on the relation between reputation and

expectations (figure 4-2). According to the results, the effects of positive WOM on

consumers’ expectations are different from the effects of negative WOM. However,

the effects did not change because of the product type. In this study, we have stressed

that the participants’ expectations arose only from the WOM we provided and had

nothing to do with factors such as previous experience, brand, and product attributes.

Thus, the participants’ expectations were affected by the content of the WOM (i.e.,

positive WOM or negative WOM) and were not related to product type.

Unlike customers’ expectations, there was no significant difference between

consumers’ actual perceived quality for positive WOM or negative WOM. However,

when specifically considering the details of product types, the results show that the

product type changes the relation between WOM and consumers’ actual perceived

quality. For experience goods, consumers’ perceived quality derived from positive

WOM is lower than that for negative WOM. For search products, the results are the

opposite. Nelson notes that the main properties of experience goods are generally not
 easy to determine except through experience. Thus, for experience goods, an objective

description is not sufficient to assist people in judging a product’s real quality,

although other users’ subjective feelings and evaluations of experience goods can be

helpful when judging quality. After receiving positive WOM, consumers will

experience a product with a positive outlook and mood. If they then discover that they

do not like the product, the ensuing gap will cause an egative relation between WOM

and perceived quality. For experience goods such as beverages, the effect is more
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

obvious. As with Wahaha Kvass, the WOM consumers receive before experiencing a

product plays an important role when customers decide whether to try it. When

customers take their first sip, they make a judgment regarding whether their perceived

quality fits their own likes and dislikes. Because the taste of the drink will curry more

favor with people who like to drink wine, the majority of consumers ’perceived

quality will not automatically be positive. Consumers who receive positive WOM will

experience a serious gap between their expectations and actual experience; their

perceived quality will be lower than that of customers who accept negative WOM.

For search products, consumers cannot have an obvious preference for perceived

quality according to their experience with a product; therefore, the perceived quality

is more likely to be consistent with their existing knowledge of products, such as

positive WOM from other people and their understanding of product attributes. Hence,

for search products, positive WOM increases customers’ perceived quality and

negative WOM decreases customers’ perceived quality.

Satisfaction is a function of expectations and perceived quality. The product type

 plays a significant regulatory role in the relation between WOM and perceived quality;

consequently, the product type also causes certain adjustments to the relation between

WOM and satisfaction. In this study, the level of satisfaction of the negative WOM

group was higher than that of the positive WOM group, although the relation changed

with different product types. For experience goods, the expectations from positive

WOM greatly improved, although perceived quality decreased, which led directly to a

decline in satisfaction. However, the results for the negative WOM group were the
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

opposite. Therefore, positive WOM reduces the degree of satisfaction, and negative

WOM improves satisfaction. For search products, however, the satisfaction of the

positive WOM group was higher than that of the negative WOM group because

positive WOM promotes expectations, although the perceived quality was created

first and the consumer was more satisfied eventually. Therefore, positive WOM

improves satisfaction, and negative WOM lowers satisfaction.

Implications and Further Research

The results of this study have implications for academia and management.

Academically, this study establishes a causal relation between WOM and customer

satisfaction, expanding the research on the relation between WOM and satisfaction. In

addition, we explore the internal mechanism of the effect of WOM on satisfaction.

Consumers’ expectations are positively influenced by WOM, and WOM has a

negative effect on consumer satisfaction, which is moderated by product type.

From a managerial perspective, these findings imply that the spread of positive

WOM can improve enterprise brand image. From a sales perspective, positive WOM
 can promote consumers’ purchase intentions and increase purchasing behavior (East

et al.,2008;Zhang, Chen, & Yang, 2012; Rodo, Leti Cia, & BeLén, 2013).However,

from a service perspective, the promotion of WOM can also create certain service

risks. When a consumer has higher expectations of a product or service and the

product or service’s actual performance does not meet expectations, the satisfaction of

the consumer may be reduced. This is because when consumers receive positive

WOM, their tolerance for a product/service domain is narrowed, and the perceived
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

quality of the product/service is reduced. Rather than tolerate the product, the

consumer will be more negative (Zeithamal, Bitner, &Gremler,2008). In addition, the

spread of negative WOM is not always bad for business. Appropriate negative WOM

can reduce consumer expectations, creating greater tolerance and rendering it easier to

improve consumer satisfaction. Some studies, however, have shown that too much

negative information may reduce consumers’ perceived quality and produce a neglect

effect, particularly when the consumer knows little about a certain product/service

(Boulding, Kalar, Staelin & Zeithaml, 1993).This type of situation will change after

subdividing the product type. Although risks and opportunities still exist for

experience goods, for search products, we should focus on the spread of positive

WOM and reduce the spread of negative WOM. This finding reminds us that we

should create different strategies according to different types of products.

Although this study has certain research and management implications, some

limitations remain. One of the limitations of this study is that WOM is a binary

variable. Future studies should quantify WOM and measure it as a continuous

 variable to better grasp the content and intensity of WOM information and to further

explore the relations between WOM and expectations, WOM and perceived quality,

and WOM and satisfaction. Another limitation is that our research only studied the

effect of product type as a regulating variable and showed that the product type has a

significant regulating effect. We did not study the role of product type indepth. We can

separate product types in the next step of this research and investigate different types

of products and how WOM specifically affects satisfaction. In addition, the perceived
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

quality and satisfaction examined in this study are product or service experiences of a

particular transaction rather than overall perceived quality and satisfaction. Data were

collected by an experiment, and the reliability and validity met the requirements.

However, acquiring related data from an extracting database would enhance the

external validity of the results of the study.

Acknowledgement
This work was partially supported by NSFC (71371034 and 71372194), the

Program for NCET (NCET-13-0687), the National Basic Research Program of China

(2012CB315805), and the Youth Research and Innovation Program in Beijing

University of Posts and Telecommunications (2014ZD02).

：
References：

Anderson,E.W. (1998), “Customer satisfaction and word of mouth”, Journal of

Service Research, Vol.8 No.1, pp. 15-17.

Anderson,E.W., Fornell,C. and Lehmann,D.R. (1994),“Customer satisfaction, market

share, and profitability: findings from Sweden”, The Journal of Marketing, pp. 53-66.
 Bansal,H.S. and Voyer,P.A. (2000), “Word-of-mouth processes within a services

purchase decision context”,Journal of service research, Vol.3 No. 2, pp.166-177.

Baron,R.M. and Kenny,D.A. (1986), “The moderator–mediator variable distinction in

social psychological research: conceptual, strategic, and statistical considerations”,

Journal of personality and social psychology, Vol.51 No.6, pp. 11-73.

Bei,T.L. Chen,I.Y. E. and Widdows,R. (2004), “Consumers’ online information search

behavior and the phenomenon of search vs. experience products”, Journal of Family
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

and Economic Issues, Vol.25 No. 4, pp. 449-467

Bickart,B. and Robert,M. S.( 2001), “Internet forums as influential sources of

consumer information”, Journal of Interactive Marketing, Vol.15 No. 3, pp. 31–40.

Blackwell,R.D., Miniard,P.W. and Engle,J.F. (2001), Consumer Behavior. New York:

Harcourt.

Brown,J.J. and Reingen,P.H. (1987), “Social ties and word-of-mouth referral

behavior”, Journal of Consumer Research, Vol.16 No.2, pp. 350-362.

Buttle,F.A. (1998), “Word of mouth: understanding and managing referral

marketing”,Journal of Strategic Marketing, Vol.6 No.3, pp. 241-254.

Chen,Y. Liu,Y. and Zhang,J. (2012), “When do third-party product reviews affect firm

value and what can firms do? The case of media critics and professional movie

reviews”, Journal of Marketing, Vol.76 No. 2, pp. 116-134.

Cohen,J. and Cohen,P. (1983), Applied Multiple Regression/ Correlation Analysis for

The Behavior Sciences (2nded.), NJ: Erlbaum.

Dong,D.H. and Liu,Y. (2012), “Contrast analysis of word-of-mouth, internet

 word-of-mouth and word-of-mouse”, Chinese Journal of Management, Vol.9 No. 3,

pp. 428-436

East,R. Hammond,K. and Lomax,W. (2008), “Measuring the impact of positive and

negative word of mouth on brand purchase probability”, International Journal of

Research in Marketing, Vol.25 No.3, pp. 215-224.

Fornell,C., Johnson,M.D., Anderson,E.W., Jaesung,C. and Bryant,B.E. (1996), “The

American customer satisfaction index: nature, purpose, and findings”, Journal of

Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

Marketing, Vol.60 No.4, pp. 7-18

Jochen,W., John,E.G. and Bateson. (1995), “An experimental investigation of halo

effects in satisfaction measures of service attributes”, International Journal of Service

Industry Management, Vol.6 No. 3, pp. 84-102

Johnson,M.D. and Fornell,C. (1991), “A framework for comparing customer

satisfaction across individuals and product categories", Journal of Consumer

Research, Vol.12, pp. 267-286

Jonathan,B. (2007), “The sales effect of word of mouth: a model for creative goods

and estimates for novels”, Journal of Cultural Economics, Vol.31 No. 1, pp. 5-23.

Keil,M., Tan,B.C.Y., Wei,K.K., Saarinen,T., Tuunainen,V. A. and Wassenaar (2000),

“A cross-cultural study on escalation of commitment behavior in software projects”,

MIS Quarterly, Vol.24 No. 2, pp. 299–325.

Klein,L.R. (1998), “Evaluating the potential of interactive media through a new lens:

search versus experience goods”,Journal of Business Research, Vol.41 No. 4, pp.

195-203.
 Lemon,K.N. White,T.B. and Winer,R.S. (2002), “Dynamic customer relationship

management: incorporating future considerations into the service retention decision”,

The Journal of Marketing, Vol. 2, pp. 1-14.

Li,Z.W. and Zhang,Y.H. (2013), “The differences of online reviews between product

and searchable”, Modern Management Science, Vol.8, pp. 23-30.

Li.D.J. and Jin,Y.H. (2005), “Empieical study of the consumer instrumentality of

WOM information source in China”, Chinese Journal of Marketing, Vol.2 No. 1, pp.
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

90-97.

Luo,X.M. and Christian,H. (2007), “Neglected outcomes of customer satisfaction”,

Journal of Marketing, Vol.71 No. 2, pp. 133-149.

Michael,D. Hartlineand Keith C. J. (1996), “Employee performance cues in a hotel

service environment: influence on perceived service quality, value, and

word-of-mouth intentions”, Journal of Business Research, Vol.35 No. 3, pp. 207–215.

Mitra,K., Reiss,M.C. and Capella,L.M. (2004), “An examination of perceived risk,

information search and behavioral intentions in search, experience and credence

services”, Journal of Services Marketing, Vol.13 No.3, pp. 208-228.

Nan,H., Pavlou,P.A. and Zhang J. (2006), “Can online reviews reveal a product's true

quality?: empirical findings and analytical modeling of online word-of-mouth

communication”, Proceeding EC '06 Proceedings of the 7th ACM conference on

Electronic commerce.

Nelson,P. (1970), “Information and consumer behavior”. Journal of Political

Economy, Vol.78 No. 2, pp. 311-329.

 Nelson,P. (1974), “Advertising as information”, Journal of Political Economy, Vol.82

No.5, pp.729-754.

NisbettRichard,E.,Wilson and Timothy,D. (1977), “Telling more than we can know:

verbal reports on mental processes”, Psychological Review, Vol.84 No.3, pp. 231-259.

Perdue,B.C. and Summers,J.O. (1986), “Checking the success of manipulations in

marketing experiments”, Journal of Marketing Research, Vol.3, pp. 317-326.

Ranaweera Chatura Prabhu and Jaideep (2003), “On the relative importance of
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

customer satisfaction and trust as determinants of customer retention and positive

word of mouth”, Journal of Targeting, Measurement and Analysis for Marketing,

Vol.12 No. 1, pp. 82-90.

Ren,X.Y. (2012), The Research of Improving the Customer Satisfaction Through Word

of Mouth Marketing, Beijing: Capital University of Economics and Business Press.

Richins,M. L., Kinnear,T.C. and Arbor,A. (1984), “Word of mouth communication as

negative information”. Association for Consumer Research, Vol.11, pp. 697-702.

Robert,A. and Westbrook. (1987), “Product/Consumption-based affective responses

and post purchase processes”, Journal of Marketing Research, Vol.24 No.3, pp.

258-270.

Sandra,P. Shainesh,G., Marcin K.,Randall,M. S.and Fernando,R. J. (2010),

“Cross-cultural differences in the effect of received word-of-mouth referral in

relational service”, Journal of International Marketing, Vol.18 No.3, pp. 62–80.

Stock,J.R. and Zinsner,P.H. (1987), “The industrial purchase decision for professional

services”, Journal of Business Research, Vol.15, pp. 1–16.

 Sun, Youn,S. ,Wu,G.H. and Kuntarapom.M. (2006), “Online word-of-mouth(or

mouse)：an exploration of its antecedents and consequences”, Journal of Internet

Mediated Communication, Vol.11 No.1, pp.1104-1127.

Sun,C.H. and Liu,Y.Z. (2009), “The Effect of e-WOM Oil Consumers Perceived

quality of Information Usefulness”, Journal of Intelligence, Vol.28 No.10, pp. 51-54.

Swan, J. E. and Oliver, R. L. (1989), “Postpurchase communications by consumers”,

Journal of Retailing, Vol.65 No.4, pp. 516-533.

Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

Tax,S.S. Chandrashekaran,M. and Christiansen,T. (1993), “Word-of-mouth in

consumer decision-making: an agenda for research”, Journal of Customer Satisfaction,

Dissatisfaction& Complaining Behavior, Vol.6, pp. 75-80.

Thorsten, H.T. and Walsh, G. (2004), “Electronic word of mouth: motives for and

consequences of reading customer articulations on the internet”, International Journal

of Electronic Commerce, Vol.8 No. 2, pp. 51-74.

Valarie,A. Z., Leonard,L. B. and Parasuraman,A. (1993), “The nature and

determinants of customer expectations of service”, Journal of the Academy of

Marketing Science, Vol.21 No. 2, pp. 1-12.

Valarie,A. Z., Parasuraman,A. and Leonard,L. B. (1985), “Problems and strategies in

services marketing”, Journal of Marketing, Vol.49, pp. 33-46.

Valarie.A.Z., Mary,J.B. and Dwayne,G.D.(2008), Services Marketing: Integrating

Cusstomer Foucs Across the Firm, Beijing: China Machine Press.

Vázquez, C. R., Suárez Á. L., Río L. A.B. (2013), “The word of mouth dynamic: how

positive (and negative) WOM drives purchase probability: an analysis of interpersonal

 and non-interpersonal factors”,Journal of Advertising Research, Vol.3 No. 2, pp.

24-38.

Westbrook,R. A. Richard and Oliver,L.(1981), “Sources of consumer satisfaction with

retail outlets”, Journal of Retailing, Vol.1 No. 3, pp. 68-85.

Whyte,W. H.and Fortune (1954), “The web of word of mouth”, Journal of Marketing,

Vol.50 No.11, pp. 140-143.

Wilbur,S. and William,E.P.(2008), Men, Women, Messages, and

Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

Media:Understanding Human Communication, Beijing: China Renmin University

Press.

Xie,J.R. (2007), “The influence of psychological effect on communication”, Journal

of Higher Correspondence education (Philosophy and Social Sciences), Vol.20 No. 8,

pp. 32-39.

Yong,L. (2006), “Word of mouth for movies: its dynamics and impact on box office

revenue”, Journal of Marketing, Vol.4 No. 2, pp. 74-89.

Zeithaml,V. and Yi,Y. (1991), “A critical review of customer satisfaction”, Review of

Marketing. Chicago, American Marketing Association.

Zhang,Q. ,Chen,D. and Yang,N. (2012), “The influence of positive iwom on consumer

behavior——college student”, China Science and Technology Review, Vol.3, pp.

284-284.
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

Figure 1．
FIGURES

．Research Framework
 Figure2. Results of Model Estimation-Positive WOM.
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

Note:*p<0.05;**p<0.01;***p<0.001; n.s: not significant

 Figure3. Results of Model Estimation-Negative WOM.
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

Note:*p<0.05;**p<0.01;***p<0.001; n.s: not significant

 Figure4. The moderating effect
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

4-1 the moderating effect of the source of WOM 4-2 the moderating effect of the product type
 TABLES

Table1. The Sample Size

WOM Source of WOM Product type Perform quantity Effective quantity

Experience goods 23 22
Online
Search goods 24 21
Positive
Experience goods 24 20
offline
Search goods 10 7
Experience goods 23 23
Online
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

Search goods 24 24
Negative
Experience goods 21 21
offline
Search goods 11 9
Total 160 147
 Table2. The reliability and confirmatory factor analysis validity

Factors Items CA Loading AVE

I believe that the product can

basic satisfy my demand for .909
drinks/mobile phone
Expecta
Expectation
ctation 0.885 0.81
I believe the product is reliable .863

I believe the overall quality of the

.919
product is good
The product can basic satisfy my
.919
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)

demand for drinks/mobile

Perceived
The product is deserve to be
0.894 .792 0.75
reliable
quality
The overall quality of the product
.885
is good
The quality of the product is short
.876
of my expectations
The product quality is the same
Satisfaction 0.886 .935 0.74
with my ideal

On the whole ,I am satisfied .769

Note. SD=standard deviation; CA = Cranach’s alpha; AVE=average variance extracted.

*p < .05. **p < .01. *** p < .000 (two-tailed).

 Table3. The Effect of WOM on the Dependent Measures

（SD）
M（ ）of Dependent Measures
Expectation*** Perceived qualityns Satisfaction**
Positive WOM 3.59（0.63） 3.05（1.01） 2.33（0.94）
Negative WOM 2.23（0.87） 3.08（1.05） 2.75（0.79）
Note:*p<0.05;**p<0.01;***p<0.001; ns: no significant
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)
Downloaded by University of Pittsburgh At 08:07 11 March 2016 (PT)