WI

1

  

WI

312% WWNWIllilllHHWIW:

  

 

   
    

LIBRARY

MIChIQaH State
niversity

Z 03"]

 

  

I

This is to certify that the
thesis entitled

COMPUTER-MEDIATED PERSUASION IN ONLINE
REVIEWS: STATISTICAL VERSUS NARRATIVE EVIDENCE

presented by

SEOYEON HONG

has been accepted towards fulfillment
of the requirements for the

M. A. degree in PUBLIC RELATIONS

 

 

/%Z~W

 

Major'Professor’s Signature

OB/ILl/éloo?

Date

MSU is an Affirmative Action/Equal Opportunity Employer

.o—.---.-.-.-u-o--o--n-c---n-o-u---ugc—n—-- —

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K'lPrqlAcc8Pres/ClRC/DateDue.indd

 

COMPUTER-MEDIATED PERSUASION IN ONLINE REVIEWS:
STATISTICAL VERSUS NARRATIVE EVIDENCE

By

Seoyeon Hong

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Public Relations

2009

ABSTRACT

COMPUTER-MEDIATED PERSUASION IN ONLINE REVIEWS:
STATISTICAL VERSUS NARRATIVE EVIDENCE

By:

Seoyeon Hong

The current study examined persuasive effects of online reviews about a product

considering summary of statistical ratings as a type of statistical evidence and consumer

 

comments as a type of narrative evidence. Study 1 used a (statistical and narrative i
review) X 2 (negative and positive review) between subject design. Study 2 used a 2

(statistical review valence: positive and negative) X 2 (narrative review valence: positive

and negative) between subject design. In both studies, the effects of online reviews were

examined on the credibility perception of the reviews, attitudes about a product, and

intention to purchase the product. Findings indicated that statistical negative reviews

were perceived to be the most credible. On the other hand, statistical reviews and

narrative reviews did not differ in affecting attitudes about the product and intention to

purchase the product. Additionally, the current study found that the vividness of

individuals’ perceptions of each review type and valence had varying effects on review

credibility, attitudes about the product, and intention to purchase the product.

This thesis is dedicated to my families: mother, Soonsup Lee, my father, Sanggi Hong,

my aunts, Hiosup Lee and Yangsup Lee, and my grandmother, Youngsoon Kim who

have given endless love, support, and encouragement to me

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LIST OF TABLES v
CHAPTER I: INTRODUCTION 1
CHAPTER 2: LITERATURE REVIEWS 3
CHAPTER 3: THEORETICAL BACKGROUND 8
CHAPTER 4: STUDY 1 15
STUDY 1 METHOD 15
STUDY 1 RESULTS 19
CHAPTER 5: STUDY 2 27
STUDY 2 METHOD 27
STUDY 2 RESULTS 30
CHAPTER 6: SUMMERY OF FINDINGS - 35
CHAPTER 7: DISCUSSION 38
CHAPTER 8: CONCLUSION 44
APPENDICES 50
APPENDIX A. Background Information 50
APPENDIX B. 2 x 2 design 52
APPENDIX C. Vividness / Credibility / Attitudes about a Product / Purchase
Intention 56

REFERENCE A ' 61

 

 

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

LIST OF TABLES

 

 

 

 

Demographic Information - Study 1 45
Online shepping Tendency - Study 1 46
Demographic Information - Study 2 47
Online shopping Tendency - Study 2 48
Comparisons of Experimental Conditions in Study 1 and Study 2 ------------- 49

INTRODUCTION

Word of mouth (WOM), as a type of interpersonal communication, is defined as
all forms of information exchanged among people regarding products, or services (Bass,
2004; Biyalogorsky, Gersyner, & Libai, 2001 ). It can be simply described as buyers
saying to each other “if you buy, I will buy” (Arndt, 1967a, p.295). People often follow
their peer group’s opinions because they are concerned about what others may think of
them, how others react to their behavior, and how credible the information they obtain
fiom others is (Arndt, 1967b).

WOM is a persuasive and distinctive marketing tactic because of its informal and
non-commercial characteristics. Brooks (1957) claimed that WOM is a key factor in
interpersonal and social influence networks and is powerful and relevant to the difiusion
of information. Consumers often have difficulty in evaluating products because of their
limited sphere of experience, and they perceive a high risk in purchasing (Heywood,
1989). Therefore, they rely on the information fi'om previous users of the product to
mitigate the risk. Walker (1995) demonstrated that more than 40% of Americans seek
others’ opinions when they need shopping advice.

WOM recommendations are perceived to be more credible (Dellarocas, 2003) and
receive more attention than other types of recommendation (Amdt, 1967; Ditcher, 1966;
Silverman, 1997). Because WOM messages originate from “people like me” who have no
personal interest in recommending a particular brand or a product, WOM is perceived to
be unprejudiced (Allsop, Bassett, & Hoskins, 2007). As a result, people tend to rely
heavily on WOM when they make purchasing decisions (Davis, Guiltman, & Jones,

1979; Richins & Root-Shaffer, 1988).

After the earliest research of Katz and Lazarfeld (1955) that claimed a sigrificant
role for WOM as a personal influence, the importance of WOM has been examined with
various products or services such as searching for a physician (Coleman, Katz, & Menzel
1957; F eldman & Spencer, 1965), new fabric (Beal & Rogers, 1957 ), information about
farming practice (Katz, 1961), razor blades (Sheth, 1971), air conditioners ONhyte, 1954),
and automobiles (Newman & Staelin, 1972). In particular, WOM is considered a major
drive in the diffusion of information (Brown & Reingen, 1987) due to its influence on a
person’s decision to adopt an innovation (Rogers, 1995; Williams, Strover, & Grant
1994). Considering the significant influence of WOM on purchasing, WOM is “the most

powerful force in the marketplace” (Silverman, 2001, p. 47).

LITERATURE REVIEW

Traditional WOM versus Electronic WOM

Since the appearance of the Internet, a number of scholars have underlined the
role of WOM in the online sphere (e. g., Goldenberg, Libai & Muller, 2001; Hennig-
Thurau et al., 2004; C. Park & Lee, 2009; D.-H. Park & Lee, 2008). Electronic WOM
(eWOM) refers to “any positive or negative statement made by potential, actual, or
former customers about a product or company, which is made available to a multitude of
people and institutions via the Internet” (Hennig—Thurau et al., 2004, p.39). Although
Hennig-Thurau et a]. (2004) claimed that the concepts of eWOM and traditional WOM
are similar, when it comes to speed, convenience, one-to-many reach and the lack of
pressure in face-to-face communication, electronic WOM is more influential than
traditional WOM (Chatterjee, 2001; C. Park & Lee, 2009; Phelps et al., 2004). Dellarocas
(2002) also claimed that eWOM is different from traditional WOM in its bi-
communicational possibilities, low cost when conducted via the Internet, the company’s
ability to control and monitor the operation, and new potential in the online situations.
Knowing the importance of eWOM, companies are currently putting forth enormous
efforts to initiate WOM online (Schwartz, 1998).

Researchers have identified five different characteristics useful in distinguishing
eWOM fiom traditional WOM (Chatterjee, 2001; Chen & Xie, 2008; Hennig-Thurau et
al., 2004; Sun, Youn, Wu, & Kuntarapom, 2006). The first characteristic is the endless
boundary: eWOM communicators can reach far beyond geographical limitations, because
consumers all over the world are accessible via the Internet. The second is the written

sphere: unlike traditional WOM which consists of spoken word exchanges, electronic

WOM takes place with written words. This makes eWOM more prevalent than traditional
WOM, because the written word is more logical than the spoken. The third characteristic
emphasizes the communication with anonymous people: consumers share their opinions
with other anonymous consumers on the Internet, while traditional WOM occurs between
friends or relatives. The fourth is directness: eWOM allows a seller to effectively and
directly initiate and broadcast consumer reviews with the help of advancement in
technologies. The last characteristic is volume: information available online is much more
voluminous in quantity compared to that of traditional WOM.

On-Line Consumer Review as a Tool of eWOM

 

With the growth of online commerce, seeking online opinions of other
consumers is increasingly becoming a part of purchasing behavior (Pitta & Fowler, 2005).
Technological developments in web sites provide unprecedented opportunities for
customers to express their opinions about companies, products, and services in the form
of online reviews (Evans, Wedande, Ralston, & Van't Hul, 2001). Online consumer
reviews have been treated as special type of WOM in many studies (e. g., Chatterjee,
2001; Chen & Xie, 2008; Godes & Mayzlin, 2004; Hu, Liu, & Zhang, 2008; D.-H. Park
& Lee, 2008). Chen and Xie (2008) discussed the implications of online customer
reviews for marketing strategies and recommended that the sellers need to develop
unique response system in online reviews. The study of DH. Park and Lee (2008)
reported that the number of reviews affects the perceived review informativeness and
popularity of the product. Online customer reviews significantly impact both companies
and customers. For example, online reviews for movies significantly affect the revenue of

the weekly box office (Liu, 2006). Out of consumers who visit online malls, 50%

consider reviews an important factor in their decision (Piller, 1999). Chevalier and
Mayzlin (2006) found that online book reviews are integral sources in book sales.
Knowing the significant impact of customer reviews, companies such as eBay.com and
Arnazon.com take action to increase the benefits of using online reviews (Melink & Alm,
2002). Online reviews can provide indirect experience before consumers make a purchase
decision (Park, Lee, & Han, 2007) and work as free ‘sales assistance’ to the e-consumers
utilizing them as a new part of marketing communications (Chen & Xie, 2008;
Wernerfelt, 1994). Online consumers utilize online reviews in order to learn about
products and reduce the uncertainty risks and transaction cost (Hu, Liu, & Zhang, 2008;
D.-H. Park & Lee, 2008).

Consumers frequently use online review that consumer- generated information is
more credible than seller-created information (Wilson & Sherrell, 1993). Also, consumer-
generated information is more interesting to the online consumer compared to marketer-
generated information (Bickart & Schindler, 2001). People consider consumer reviews
created by users to be for the benefit of other users, so their information will contain
more possible usages (Chen & Xie, 2008). Additionally, online reviews enable buyers to
overcome the limitations of an online store; it is impossible to see, touch and smell the
product (Chatterjee, 2001; Chen & Xie, 2008; Park, Lee, & Han, 2007). In other words,
online review meets consumers’ need as they are looking for recommendations about
purchasing a product. Therefore, the online review as a form of eWOM communication
has become a crucial source for both consumers and marketers.

The feedback system that eBay employs is an excellent example of online

consumer review (Chevalier & Mayzlin, 2006). The feedback system allows consumers

to acquire verifiable information about sellers who may be unknown in cyberspace.
Sellers can build their reputation by increasing the number of transactions they make on
eBay (Resnick & Zeckhauser, 2002). A number of studies have examined the efficiency
of the eBay feedback system. For example, Resnick, Zeckhauser, Swason, and Lockwood
(2006) showed that users have stronger willingness to purchase the product from a seller
with a high reputation rather than a seller with a newly created account. Additionally,
positive feedback increased the price of products (Dewan & Hsu, 2004; Kalyanarn &
McIntyre, 2001; Mcdonald & Slawson Jr., 2002; Melink & Alm, 2002) while negative
feedback decreased the price (Kalyanam & McIntyre, 2001; Melink & Alm, 2002) and
probability of sale (Melink & Alm, 2002).
Computer-Mediated Persuasion

O’Keefe (2004) pointed out that despite today’s widespread use of computers,
few studies investigated the aspects of “computer mediated persuasion.” Moreover,
researchers have emphasized the necessity of considering the influence of computer
mediated persuasion on the use and effectiveness of evidence (Reynolds & Reynolds,
2002; Seiter & Seiter, 2005). Therefore, the current research aims to fill the gap between
persuasion research in offline and online environments and improve understanding of the
persuasion effects of online evidence. Drawing from the literature on WOM and
persuasion, this study seeks to build a connection between them in the context of online
reviews. The present study compares the two types of online reviews: narrative and
statistical evidence.

Online reviews can be classified into two types of feedback. One is a statistical

feedback (e.g., 999 customers out of 1000 satisfied with the product) and another is a

narrative feedback (e.g., I strongly recommend this camera. When I went to Sea World at
night, it was easy to take a picture in the darkness with its amazing flash function). The
feedback system in eBay provides both types of feedback on the same web page.
Inconsistencies between the statistical and the narrative feedback can occur. Even when
the statistical feedback is very positive, the narrative feedback can be negative for the
same product. Then, one may ask which kind of feedback is more likely to influence

online consumers' attitudes about a product.

THEORETICAL BACKGROUND
Persuasion and Evidence
McCroskey (1969) defined evidence as “factual statements originating fi'om a

source other than the speaker, objects not created by the speaker, and opinions of persons
other than the speaker that are offered in support of the speaker’s claims” (p. 170).
Evidence includes relevant facts, opinions, and information that support the persuader’s
assertion (O’Keefe, 2002). In general, evidence is employed to increase the
persuasiveness of the message and provide argumentative support for the message.
Evidence consists of the facts and opinions creating a logical basis for the argument
(Abernathy, 1964; Cohen & Nagel, 1962). Past research has shown that use of evidence
is more persuasive when message recipients have personal involvement in an issue
(Cacioppo, Petty, Kao, & Rodriguez, 1986; Petty & Cacioppo, 1984; Petty, Cacioppo, &
Goldman, 1981; Reynolds & Reynolds, 2002; Seiter & Seiter, 2005).
Statistical versus Narrative Evidence

The current paper examines two types of evidence: statistical and narrative
evidence. Statistical evidence is one type of factual information (Baesler & Burgoon,
1994; Reinard, 1988), includes “quantified descriptions of events, persons, places, or
other phenomena” (Church & Wulbanks, 1986, p. 108), and is characterized as proof in
the form of summary information across a number of cases (Allen & Preiss, 1997).
Dickson (1982) considered “a rate, frequency, proportion, percentage, probability,
average, median, or other statistical parameter” as statistical information (p. 399).

Persuasiveness of statistical evidence can depend on the number of observations, because

a larger number of cases can indicate stronger objectivity. Moreover, statistical evidence
has a direct impact on judgments and an indirect impact on attitudes (Boster et a1, 2000).

Narrative evidence is a message that presents information in a personal format as
opposed to a statistical format (Ameson & Query, 2001; Ellinson & Buzzanell, 1999;
Vanderford, Smith, & Harris, 1992). Narrative evidence is defined as the “use of case
stories or examples to indicate that the conclusion offered by the communicator is true”
(Allen & Preiss, 1997). Narrative evidence includes testimonial assertions which Reinard
(1988) referred to as judgments and opinions fi'om others, and also includes what
O'Keefe (2002) refers to as description and information in detail about an instance or an
event. Generally, it focuses on elaborating the example of an event and providing
appealing detail, characters, and some plot (Greene & Brinn, 2003).

A number of studies have been conducted to examine the effects of evidence.
Particularly, the comparison between narrative and statistical evidence has received
attention from researchers. Previous studies stated that narrative evidence is more
persuasive (Kazoleas, 1993; Kahneman & Trversky, 1973; Nisbett & Ross, 1980; Nisbett
& Borgida, 1975; Taylor & Tompson, 1982). On the other hand, Reinard (1988) showed
that both statistical and narrative evidence are equally persuasive and memorable. Also,
other studies reported that both statistical and narrative evidence produce persuasive
effects equally and reported no difference between the two types (Iyengar & Kinder,
1987; Nadler, 1983; Reyna, Woodruff, & Brainerd, 1987). Therefore, as O'Keefe (2002)
pointed out, it is unclear why and how one type of evidence is more effective than the

other.

Putting aside the debates, a number of studies have continually emphasized the
superior effect of statistical evidence. A meta-analysis by Allen and Preiss (1997) showed
statistical evidence to be slightly more effective. Correspondingly, statistical messages
are found to be more effective for health information than narrative information but both
statistical and narrative messages are better than none at all (Greene & Brinn, 2003).
Kopfrnan, Smith, Yun, and Hodges (1998) noted that messages containing statistical
evidence produced greater cognitive reactions while narrative evidence produced greater
affective reactions. Moreover, Baesler and Burgoon (1994) found that both forms of
evidence were initially persuasive when compared to no evidence at all; however;
statistical evidence was more persuasive in the long term. Baesler and Burgoon (1994)
pointed out the fact that only a few studies on statistical versus narrative evidence used
representative forms of narrative evidence, and they criticized the absence of studies
addressing vividness.

Vividness is “information that’s emotionally interesting, concrete and imagery
provoking, proximate in a sensory, temporal, or spatial way” (N isbett &Ross, 1980, p.
45). Narrative evidence is easier to imagine than statistical evidence because narrative
evidence has vividness as a factor that bolsters the effect of persuasion. On the other hand,
statistical evidence, considered to be more abstract than narrative, fails to facilitate
involvement, increase interest, or engage attention. Baesler and Burgoon (1994)
controlled vividness when comparing the two types of evidence and found that with
vividness, statistical evidence was generally more persuasive and produced immediate,
moderate term (48 hours) and long term (one week) persuasive effects. Considering the

potential role of vividness as a confounding factor, it seems important to consider and

10

control the effect of vividness in order to conduct a fair comparison of statistical and
narrative evidence. It is expected that statistical evidence is more compelling when
vividness is controlled.
Persuasive Effects

In the current study, persuasive effects of online reviews will be assessed in three
ways: credibility of reviews as evidence, attitudes about a product, and intention to
purchase a product. Evidence can affect perceived credibility which will eventually
increase persuasive effects (Reinard, 1988). Perceived credibility refers to “the judgments
made by a perceiver (e.g., a message recipient) concerning the believability of a
communicator” (O’keefe, 2002, p. 181). Although credibility often pertains to an
information source or sender, online reviews as a form of eWOM are provided by
anonymous and numerous people, rather than a particular one person. Thus, credibility
can be also relevant for online. reviews. For example, F lanagin and Metzger (2000)
examined credibility of Internet information. When individuals see online reviews of a
product, the extent to which they trust the reviews may vary. Online reviews can be
persuasive in terms of their effects on individuals' evaluations of the product and
intentions to purchase the product to which online reviews pertain. Attitudes about a
product and purchasing intention have been a popular outcome of interests in previous
studies examining the influence of WOM (e.g., Bansal & Voyer, 2000; C. Park & Lee,
2009; Sen & Lennan, 2007; Wangenheim & Bayou, 2004; Ying & Chung, 2006). It is
expected that different types of evidence can be perceived more or less trustworthy and

believable, and can result in varying levels of attitudes and purchasing intentions.

11

Hypothesis l-a: Individuals will perceive online statistical reviews as more
credible than online narrative reviews when vividness is controlled.

Hypothesis l-b: Individuals will indicate more positive attitudes about a product
after seeing online statistical reviews than online narrative reviews when
vividness is controlled.

Hypothesis 1-c: Individuals will indicate stronger intention to purchase a product
after seeing online statistical reviews than online narrative reviews when
vividness is controlled.

Given that evidence means facts or opinions presented as proof for an assertion
(Reynolds & Reynolds, 2002), evidence can also vary as negative or positive for a
product at issue. Both positive and negative information can influence the attitude
change, but the impact of negative messages on consumers is considerably more
significant (Laczniak, DeCarlo, & Ramaswami, 2001). Satisfied consumers rarely
mention their experience, but dissatisfied customers tell everyone (Holmes & Lett 1977;
Swan & Oliver 1989). Additionally, research on WOM shows that negative WOM has
larger impacts on purchasing decisions and on consumer behavior than positive WOM
(Herr, Kardes, & Kim, 1991).

The valence of evidence (i.e., negative and positive evidence) can be differentially
persuasive when individuals consider purchasing or not purchasing a product. Of the past
studies of evidence, most studies used negative messages to show the effectiveness of the
evidence regardless of the comparison results. Only a small number of studies used
positive messages (e.g., Hamill, Wilson, & Nisbett, 1980; Hoeken, 2001; Kahneman, &

Tversky, 1973 ; Koballa, 1986). For example, the past studies comparing statistical and

12

narrative evidence used negative evidence or messages about topics such as crime rates
(Caroll, I97 7), car accidents and safety belts (Kazoleas, 1993), electronic accidents
(Wells & Harvey, 1977), car accidents and organ donation (Kopfrnan et al., 1998),
pollution (N adler, 1982), failed refrigerators (Dickson, 1982), divorce and juvenile
delinquency (Baesler & Burgoon, 1994), alcohol education messages (Slater & Rouner,
1996), and skin cancer (Greene & Brinn, 2003). Of the studies using negative evidence,
seven studies concluded that statistical evidence has greater persuasive effects. Thus, it is
possible that the difference between statistical and narrative evidence can be greater when
the evidence is negative toward a product than when it is positive. Thus, the following
hypothesis is advanced:

Hypothesis 2: The difference between statistical online review and narrative
online review containing negative information will be larger than the
difference between statistical and narrative online review containing
positive information.

Overview of Study 1 and Study 2

Two studies were conducted to examine the persuasive effects of review type and
review valence. Study 2 replicated study 1 with only one difference. Study 1 used a 2
(review type: statistical and narrative review) X 2 (review valence: negative and positive
review) between subject design. Study 2 used a 2 (statistical review valence: positive and
negative) X 2 (narrative review valence: positive and negative) between subject design.
For example, participants in one of four experimental conditions of study 1 read positive
statistical review, whereas participants in one of four experimental conditions read

statistical positive review as well as narrative positive review. After reading reviews

13

about a product (a digital camera), participants in both studies indicated their credibility

perception of reviews, attitudes about the product, and intention to purchase the product.

14

Study 1
Study 1 Method

Participants

One hundred eighty one students (68.51% women, age M = 22.35, SD = 3.09)
enrolled at Michigan State University participated in the study in exchange of course
credit. Of the participants, 70.4% were Caucasian, 8.3% were Asian or Asian American,
3.7% were Afi'ican American, 2.8% were Pacific islander, 0.9% were Hispanic, and
13.9% were other (i.e., unclassifiable, mixed, etc.). Tables 1 and 2 provide more detailed
information about the demographic characteristics and online shopping tendencies of
participants.
Design and Procedure

This study used a 2 (review type: statistical and narrative review) X 2 (review
valence: negative and positive review) between subject design. The experimental
materials included information about a product (a digital camera) and were designed
following typical online review formats commonly found on eBay.com, for example.
Statistical review included the number of ratings (e.g., positive review = 900 positive
ratings out of 1000; negative review = 900 negative ratings out of 1000). For narrative
review, examples of positive review included "Amazing quality" and "I will recommend
this to everyone." Examples of negative review included "Disappointing quality" and "I
won't recommend this to anyone." See appendix II for the experimental materials. This
study also included a control condition that provided only the product information and no

online reviews. In the control condition, participants answered questions only about their

15

attitudes towards the product and intention to purchase the product, not about review
credibility and vividness. See appendix III for all the measurement items.
Manipulation Check

Review valence. For positive and negative reviews, 10 items (a = 0.99) assessed
the valence of reviews on a 7-point scale (1 = strongly disagree, 7 = strongly agree).
Example items included, "overall reviews are positive" and "almost all of reviews are
recommending this product."

A 2 (review type: statistical and narrative review) X 2 (review valence: negative
and positive review) between-subject ANOVA was conducted on valence. Review type
did not have a significant main effect, F (1, 141) = 0.40, p = .53, 112 = .00. Narrative
review (M = 3.92, SD = 2.29) did not differ from statistical review (M = 3.85, SD = 2.04).
Review valence had a significant main effect, F (l, 141) = 923.15,p < .001, 112 = .87.
Positive review (M = 5.81, SD = 0.81) was perceived as more positive than statistical
review (M = 1.82, SD = 0.80). The interaction between review type and valence was
significant, F (1, 141) = 5.97, p = .01, n2 = .006. For negative review, statistical review
(M = 1.94, SD = 0.84) did not differ from narrative review (M = 1.70, SD = 0.74), t (68) =
1.26, p = .21. For positive review, narrative review (M = 6.02, SD = 0.71) was perceived
as more positive than narrative review (M = 5.62, SD = 0.85), t (73) = 2.22, p = .03.

For statistical review manipulation, positive review (M = 5.34, SD = 1.14) was
perceived as much more positive than negative review (M = 2.10, SD = 1.18), t (168) =
18.18, p < .001, n2 = .66. Positive review rating was significantly higher than the

midpoint (4) of the scale, t (88) = 11.05, p < .001. Negative review rating was significant

16

lower than the midpoint of the scale, indicating that negative review was perceived as
negative, t (80) = -14.49, p < .001.

For narrative review manipulation, positive review (M = 5.81, SD = 0.80) was
perceived as much more positive than negative review (M = 1.82, SD = 0.73), t (168) =
34.10, p < .001, 112 = .87. Positive review rating was significantly higher than the
midpoint (4) of the scale, t (85) = 23.07, p < .001. Negative review rating was significant
lower than the midpoint of the scale, indicating that negative review was perceived as
negative, t (83) = -25.07, p < .001.

Vividness. The vividness imagery was rated with items modified and adapted
from Nisbett and Ross (1980) and Baesler and Burgoon (1994). Each two types (personal
and general) of four items (vividness, emotiveness, concreteness, and imaginativeness)
assessed vividness on a 7-point scale (1 = not at all, 7 = very much). These eight items
for statistical vividness had a reliability of a = 0.93 and narrative vividness had a = 0.90.

A 2 (review type: statistical and narrative review) X 2 (review valence: negative
and positive review) between-subject ANOVA showed that review type had a significant
main effect, F(1, 141) = 18.74,p < .001, n2 = .12. Narrative review (M: 4.64, SD =
0.97) was considered more vivid than statistical review (M = 3.82, SD = 1.30). Review
valence did not have a significant main effect, F (l, 141) = 0.09, p = .76, n2 = .00. The
interaction between review type and valence was not significant, F (1, 141) = 1.51, p =
.12, n2 = .015.

Measurements of Dependent Variables
Review credibility. F lanagin and Metzger (2000) identified 5 items popularly used

to measure information credibility (biased-unbiased, trustworthy-untrustworthy, accurate-

17

inaccurate, believable-unbelievable, and complete-incomplete). These five items (a =
0.80) were used in the current study (e.g., "To me, overall reviews are ..... ") with a 7-point
scale (e.g., 1 = extremely unbelievable, 7 = extremely believable).

Attitudes toward the product. Attitudes were measured with three items (a = 0.97)
(good, favorable, and desirable). Participants were asked to indicate their feelings about
this product on a 7-point scale (1 = extremely bad, 7 = extremely good).

Behavioral intention to purchase the product. Purchasing intention was measured
with 5 items (a = 0.96). Example items included "If I am going to buy a new digital
camera in the future, I intend to buy this produc " and "If I am going to buy a new digital

camera in the future, I would not buy this product" (reverse-coded).

l8

Study 1 Results

Hypothesis 1

Hypothesis l-a predicted that online statistical review would be more credible
than online narrative review when vividness was controlled. The hypothesis was tested
with a 2 (review type: statistical and narrative review) X 2 (review valence: negative and
positive review) between-subject ANCOVA with vividness as a covariate. The results
showed that vividness was not significant, F (1, 140) = 0.20, p = .65, n2 = .00. Review
type had a significant main effect on review credibility, F (l , 140) = 9.32, p = .003, n2
= .06. Statistical review (M = 4.78, SD = 1.00) was perceived as more credible than
narrative review (M = 4.29, SD = 0.77). Thus, the data were consistent with Hl-a.

Additionally, review valence had a significant main effect on review credibility, F
(1, 140) = 9.43, p = .003, n2 = .06. Negative review (M = 4.78, so = 1.05) was perceived
as more credible than positive review (M = 4.33, SD = 0.74). The interaction effect of
review type by valence was not significant, F ( 1, 140) = 1.43, p = .23, 112 = .009.

Hypothesis 1-b predicted that attitudes about the product would be more positive
after seeing online statistical review than online narrative review when vividness was
controlled. The hypothesis was tested with a 2 (review type: statistical and narrative
review) X 2 (review valence: negative and positive review) between-subject AN COVA
with vividness as a covariate. The results showed that vividness was not significant, F (1,
140) = 1.89, p = .17, n2 = .004. Review type did not have a significant main effect on
attitudes about the product, F (1, 140) = 0.84, p = .36, n2 = .00. Statistical review (M =
3.64, SD = 1.72) and narrative review (M = 3.83, SD = 1.49) did not differ from one

another. Thus, the data were not consistent with H1 -b.

19

Additionally, review valence had a significant main effect on attitudes about the
product, F (1, 140) = 341 .46,p < .001, n2 = .70. Positive review (M = 5.03, SD = 0.74)
resulted in more positive attitudes about the product than negative review (M = 2.34, SD
= 1.01). The interaction effect of review type by valence was not significant, F (1, 140) =
0.81, p -—- .37, n2 = .00.

Hypothesis 1-c predicted that intention to purchase a product would be stronger
after reading online statistical review than online narrative review when vividness was
controlled. A 2 (review type: statistical and narrative review) X 2 (review valence:
negative and positive review) between-subject ANCOVA with vividness as a covariate
was conducted. The results showed that vividness was not significant, F (1, 140) = 0.07,
p = .78, n2 = .00. Review type did not have a significant main effect on behavioral
intention to purchase the product, F (1, 140) = 0.001, p = .98, n2 = .00. Statistical review
(M = 3.27, SD = 1.64) and narrative review (M = 3.27, SD = 1.44) did not differ from one
another. Thus, the data were not consistent with H 1 -c.

Additionally, review valence had a significant main effect on intention to
purchase the product, F (1, 140) = 192.95, p < .001, 112 = .57. Positive review (M = 4.40,
SD = 1.02) resulted in greater intention than negative review (M = 2.06, SD = 0.99). The
interaction effect of review type by valence was not significant, F (1, 140) = 1.29, p = .26,
n2 = .004.

Hypothesis 2
Hypothesis 2 predicted that the difference between statistical and narrative online

review containing negative information would be larger than the difference between

20

statistical and narrative online review containing positive information. The hypothesis
was tested with one way AN OVAs for each of the three dependent variables.

Review credibility. A one way AN OVA comparing the statistical negative,
narrative negative, statistical positive, and narrative positive conditions showed that the
experimental conditions had a significant main effect on credibility perception of review,
F (3, 141) = 7.59, p < .001, n2 = .14. A post hoc comparison using Tukey showed that
statistical negative review (M = 5.11, SD = 1.20) was perceived as more credible than
narrative negative review (M = 4.43, SD = 0.72) and that statistical positive review (M =
4.48, SD = 0.65) did not differ from narrative positive review (M = 4.17, SD = 0.80).
There were no significant differences among narrative negative, statistical positive, and
narrative positive reviews.

Attitudes toward the product. A one way ANOVA comparing the statistical
negative, narrative negative, statistical positive, and narrative positive review conditions
and the control condition showed that the experimental conditions had a significant main
effect on attitudes about the product, F (4, 176) = 92.71, p < .001, n2 = .68. A post hoc
comparison using Tukey showed that statistical negative review condition (M = 2.16, SD
= 1.09) and narrative negative review condition (M = 2.53, SD = 0.90) did not
significantly differ fi'om one another and that statistical positive review condition (M =
5.00, SD = 0.82) and narrative positive condition (M = 5.06, SD = 0.64) did not
significantly differ fiom one another. Statistical and narrative negative review conditions
differed significantly fiom statistical and narrative positive review conditions. The
control condition (M = 4.16, SD = 0.76) differed significantly from both negative and

positive review conditions.

21

Because statistical and narrative review conditions did not differ fiom one
another, both conditions were combined across negative and positive review conditions.
Which of negative and positive review conditions affected attitudes to a greater extent
was examined. As expected, attitudes about the product were less positive in the negative
review condition (M = 2.34, SD = 1.01) than the control condition (M = 4.16, SD = 0.76),
t (104) = 9.52, p < .001, point-biserial r = .68. Attitudes about the product were more
positive in the positive review condition (M = 5.03, SD = 0.74) than in the control
condition (M = 4.16, SD = 0.76), t (109) = 5.77, p < .001, point-biserial r = .48.
Comparing the two correlations (.68 versus .48) indicated that the extent to which the
negative review condition decreased attitudes about the product compared to the control
condition was greater than the extent to which the positive review condition increased
attitudes compared to the control condition, 2 = 2.20, p = .03, two-tailed.

Intention to purchase the product. A one way ANOVA comparing the statistical
negative, narrative negative, statistical positive, and narrative positive conditions and the
control condition showed that the experimental conditions had a significant main effect
on behavioral intention to buy the product, F (4, 176) = 48.17, p < .001, n2 = .52. A post
hoc comparison using Tukey showed that statistical negative condition (M = 1.96, SD =
0.96) and narrative negative condition (M = 2.16, SD = 1.02) did not significantly differ
fi'om one another and that statistical positive condition (M = 4.49, SD = 1.11) and
narrative positive condition (M = 4.30, SD = 0.91) did not significantly differ from one
another. Statistical and narrative negative conditions differed significantly from statistical
and narrative positive conditions. The control condition (M = 3.20, SD = 1.07) differed

significantly from both negative and positive conditions.

22

Because statistical and narrative review conditions did not differ fiom one
another, both conditions were combined across negative and positive review conditions
and comparisons were made in order to examine which of negative and positive review
conditions affected behavioral intention to a greater extent. As expected, intention to
purchase the product was weaker in the negative review condition (M = 2.06, SD = 0.99)
than in the control condition (M = 3.20, SD = 1.07), t (104) = 5.49, p < .001, r = .47.
Intention to purchase the product was higher in the positive review condition (M = 4.40,
SD = 1.02) than in the control condition (M = 3.20, SD = 1.07), t (109) = 5.49, p < .001, r
= .48. Comparing the two correlations (.47 versus .48) indicated that the extent to which
the negative review condition decreased behavioral intention compared to the control
condition did not differ from the extent to which the positive review condition increased
behavioral intention compared to the control condition, 2 = 0.09, p = .93, two-tailed.
Additional Analyses.

The aforementioned ANCOVA and ANOVA results for hypotheses 1 and 2
indicated that vividness was not a significant covariate and that inclusion or exclusion of
vividness as a covariate did not change the findings regarding the effects of evidence type
on review credibility, attitudes about the product, and intention to purchase the product.
In order to further examine the role of vividness, however, exploratory analyses were
conducted. Moderated multiple'regression analyses examined the possibility of vividness
as a moderator for the effects of review type and review valence on review credibility,
attitudes about the product, and behavioral intention to purchase the product. Each of the
two categorical variables was dummy-coded; statistical review = 0 and narrative review =

1 for review type, negative review = 0 and positive review = 1 for review valence. One

23

continuous variable, vividness, was mean-centered and multiplied by review type and by
review valence to create second-order predictors (i.e., review type X vividness and review
valence X vividness) and a third-order predictor (i.e., review type X review valence X
vividness) for testing interactions. Hierarchical multiple regression analyses were
conducted with the first-order predictors in the first block, the second-order predictors in
the second block, and the third-order predictor in the third block. For any significant
interactions, simple regression analyses were conducted to probe the interaction pattern.
Review Credibility

The overall model with all the predictors was significant, F (6, 138) = 7.49, p <
.001 , adj. R2 = .21. The first-order predictors in the first block of the regression model
explained a significant amount of variance in review credibility, F (3, 141) = 7.12, p <
.001, adj. R2 = .13. Among the first-order predictors, vividness was not significant, [3 = -
.05, t = -0.61, p = .54, sr (semipartial correlation) = -.05. On the other hand, review type
was significant, [3 = -.25, t = -2.96, p = .004, sr = -.23, indicating that statistical review
was perceived as more credible than narrative review. Review valence was significant, [3
= -.24, t = -3.10, p = .002, sr = -.24, indicating that negative review was perceived as
more credible than positive review.

The two second-order predictors accounted for a significant amount of variance in
review credibility, F change (2, 139) = 3.31, p = .04, chhange = .04. The interaction of
review type by vividness was not significant, B (unstandardized coefficient) = 0.23, t =
1.74, p = .09, sr = .13. The interaction of review valence by vividness was significant, B
= 0.23, t = 1.98, p = .049, sr = .15. Simple slope of vividness was negative for negative

review, b = -0.13, p = .22, whereas simple slope of vividness was positive for positive

24

review, b = 0.07, p = .34. This second-order interaction was qualified by a third-order
interaction, which clarified that the interaction of review type by vividness varied with
review valence.

The third-order predictor also accounted for a significant amount of variance in
review credibility, Fchange (1, 138) = 13.66, p < .001, chhange = .08. The interaction of
review type and review valence with vividness was significant, B = -0.93, t = -3.67, p <
.001, sr = —.27. For statistical review, simple slopes of vividness were negative for
negative review, b = -0.3 7, p = .01, and positive for positive review, b = 0.16, p = .049.
For narrative review, simple slopes of vividness were positive for negative review, b =
0.31, p = .009, and negative for positive review, b = -0.12, p = .44.

Attitudes toward the Product

The overall model with all the predictors was significant, F (6, 138) = 33.80, p <
.001, adj. R2 = .58. The first-order predictors in the first block of the regression model
explained a significant amount of variance in attitudes about the product, F (3, 141) =
64.80, p < .001, adj. R2 = .57. Among the first-order predictors, vividness was not
significant, [3 = .02, t = 0.41, p = .68, sr = .02. Review type was not significant, [3 = -.01, t
= -0.12, p = .91, sr = -.01. Review valence was significant, B = .76, t = 13.92, p < .001, sr
= .76, indicating that attitudes about the product were more positive in the positive review
condition than in the negative review condition. The two second-order predictors did not
account for a significant amount of variance in attitudes, F change (2, 139) = 2.60, p = .08,
chhange = .015. The third-order predictor also did not account for a significant amount of
variance in attitudes, Fchange (1, 138) = 0.11, p = .74, chhange = .00.

Behavioral Intention to Purchase the Product

25

The overall model with all the predictors was significant, F (6, 138) = 61 .91, p <
.001, adj. R2 = .72. The model with the first-order predictors was significant, F (3, 141) =
116.43, p < .001, adj. R2 = .71. Among the first-order predictors, vividness was not
significant, B = .07, t = 1.51, p = .13, sr = .07. Review type was not significant, B = .04, t
= 0.85, p = .40, sr = .04. Review valence was significant, B = .84, t = 18.53, p < .001, sr =
.84, indicating that behavioral intention to purchase the product was greater in positive
review condition than in negative review condition.

The two second-order predictors accounted for a significant amount of variance in
attitudes, Fchange (2, 139) = 3.54, p = .04, chhange = .014. The interaction of review type
by vividness was significant, B = -0.32, t = -2.43, p = .02, sr = .11. Simple slope of
vividness was positive for statistical review, b =.12, p = .21, whereas simple slope of
vividness was negative for narrative review, b = -.17, p = .15. The interaction of review
valence with vividness was not significant, B = 0.12, t = 0.97, p = .33, sr = .04. The third-
order predictor also did not account for a significant amount of variance in attitudes,

Fchange (1, 138) = 1.42. p = .24, chhange = .003.

26

Study 2

In study 1, participants read only one type of reviews, either statistical or
narrative. But because most Web sites in reality present both types of reviews, study 1
might have lacked realism. As a way to replicate study 1 and also to increase realism,
study 2 was designed using the same materials as study 1, except that participants in
study 2 read both types of reviews.

Study 2 Method

Participants

One hundred forty five students (67.59% women, age M = 22.70, SD = 3.35)
enrolled at Michigan State University participated in the study in exchange of course
credit. Of the participants, 76.5% were Caucasian, 8.2% were Asian or Asian American,
3.5% were African American, 1.2% were Hispanic, and 10.6% were other (i.e.,
unclassifiable, mixed, etc.). Tables 3 and 4 provide more detailed information about the
demographic characteristics and online shopping tendencies of participants.
Design and Procedure

This study used a 2 (statistical review valence: positive and negative) X 2
(narrative review valence: positive and negative) between subject design. Unlike study 1
where participants read either statistical or narrative review only,_all participants in study
2 read both statistical and narrative review. For example, participants in one of four
experimental conditions read statistical positive review and narrative negative review and
evaluated each review on review valence and vividness. Then, participants indicated their

credibility perception, attitudes about the product, and intention to buy the product. The

27

experimental materials and measurement items used in study 2 were identical to study 1
(see appendices A, B, and C).
Manipulation Check

Review valence. Ten items assessed the valence of reviews (or = 0.99 for statistical
review and a = 0.99 for narrative review) on a 7-point scale (1 = strongly disagree, 7 =
strongly agree). Statistical review positivity was not significantly correlated with
narrative review positivity, r (168) = .08, p = .29. For statistical review manipulation,
positive review (M = 5.34, SD = 1.14) was perceived as much more positive than
negative review (M= 2.10, SD = 1.18), t (168) = 18.18,p < .001, n2 = .66. Positive
review rating was significantly higher than the midpoint (4) of the scale, t (88) = 11.05, p
< .001. Negative review rating was significant lower than the midpoint of the scale,
indicating that negative review was perceived as negative, t (80) = -14.49, p < .001.

For narrative review manipulation, positive review (M = 5.81, SD = 0.80) was
perceived as much more positive than negative review (M = 1.82, SD = 0.73), t (168) =
34.10, p < .001, n2 = .87. Positive review rating was significantly higher than the

midpoint (4) of the scale, t (85) = 23.07, p < .001. Negative review rating was significant
0 lower than the midpoint of the scale, indicating that negative review was perceived as
negative, t (83) = -25.07, p < .001.
Vividness

For statistical vividness, the eight items had a reliability of or = 0.89 and narrative
vividness had 01 = 0.92. Items (e. g., "In general, to what extent were these messages
vividly presented?" and "To you, personally, to what extent were these messages

concretely presented?") used a 7-point scale (1 = not at all, 7 = very much). Correlation

28

between statistical review vividness and narrative review vividness was small but
significant, r (168) = .28, p < .001. T-tests were conducted to see if statistical review and
narrative review differed in vividness. The results showed that narrative review (M =
4.73, SD = 1.02) was perceived as more vivid than statistical review (M = 3.67, SD =
1.05), paired sample t (169) = 11.16, p < .001, n2 = .42. For the vividness rating of
narrative review, negative review (M = 5.06, SD = 0.88) was perceived as more vivid
than positive review (M = 4.40, SD = 1.04), independent samples t (168) = 4.46, p < .001,
n2 = .11. For vividness rating of statistical review, negative review (M = 3.72, SD = 1.08)
and positive review (M = 3.62, SD = 1.01) did not differ, independent samples t (168) =
0.65, p = .52, n2 = .00.
Measurements of Dependent Variables

Five items (a = 0.84) assessed review credibility (e.g., "To me, overall reviews
are ..... ", 1 = extremely biased, 7 = extremely unbiased). Three items (a = 0.98) assessed
attitudes about the product (e.g., 1 = extremely unfavorable, 7 = extremely favorable).
Five items (0. = 0.95) assessed intention to purchase the product (e. g., "If I am going to

buy a new digital camera in the future, I intend to buy this product").

29

Study 2 Results

Overview

Moderated multiple regression analyses was conducted to examine the persuasive
effects of statistical review valence and narrative review valence as well as to test a
possibility of vividness as a moderator for the effects of statistical review valence and
narrative review valence on review credibility, attitudes about the product, and behavioral
intention to purchase the product. Each of the two categorical variables was dummy-
coded; statistical negative review = 0 and statistical positive review = 1 for statistical
review valence, narrative negative review = 0 and narrative positive review = 1 for
narrative review valence. Two continuous variables (statistical review vividness and
narrative review vividness) were mean-centered and multiplied by statistical review
valence and by narrative review valence to create second-order predictors (i.e., statistical
review vividness X statistical review valence, narrative review vividness X narrative
review valence). Additionally, a product term of statistical review valence X narrative
review valence was created as a second-order predictor. Hierarchical multiple regression
analyses were conducted with the first-order predictors in the first block and the second-
order predictors in the second block. For any significant interactions, simple regression
analyses and post hoc comparisons were conducted to probe the interaction pattern.
Review Credibility

The overall model with all the predictors was significant, F (7, 159) = 12.48, p <
.001, adj.R2 = .33. The first-order predictor in the first block of the regression model
accounted for a significant amount of variance in review credibility, F (4, 162) = 13.74, p

<. 001, adj.R2 = .24. Among the first-order predictors, vividness of statistical review was

30

inversely related to review credibility, B = -.28, t = -4.00, p < .001, sr = -.27, and
vividness of narrative review was positively related to review credibility, B = .39, t =
5.19, p < .001, sr = .35. Valence of statistical review was also a significant predictor, B =
-.19, t = -2.73, p = .007, sr = -.19, indicating that statistical negative review was
perceived as more credible than statistical positive review. On the other hand, valence of
narrative review was not a significant predictor, B = -.12, t = -1 .71, p = .09, sr = -.12.

The second-order predictors in the second block of the regression model
explained a significant amount of variance in review credibility, F change (3, 159) = 8.31,
p < .001, chhange = .10. The interaction of narrative review valence with narrative review
vividness was not significant, B = 0.03, t = 0.20, p = .84, sr = .01. However, the
interaction of statistical review valence by statistical review vividness was significant, B
= 0.41, t = 3.49, p = .001, sr = .22. Simple slope of statistical review vividness was
negative for statistical negative review, b = -0.45, p < .001, whereas simple slope of
statistical review vividness was close to zero for statistical positive review, b = 0.01 , p =
.89. This finding indicated that for statistical negative review, the more vivid the review
was, the less credible the review was. Additionally, the interaction of statistical review
valence by narrative review valence was significant, B = 0.93, t = 3.83, p < .001, sr = .24,
indicating that statistical negative review coupled with narrative negative review (M =
5.16,, SD = 1.24) was perceived as more credible than statistical negative review coupled
with narrative positive review (M = 4.241,, SD = 0.73), statistical positive review coupled
with narrative negative review (M = 4.33b, SD = 0.79), and statistical positive review
coupled with narrative positive review (M = 4.36b, SD = 0.75).

Attitudes toward the Product

31

The overall model with all predictors was significant, F (7, 159) = 23.04, p <
.001, adj.R2 = .48. The first-order predictor in the first block of the regression model
accounted for a significant amount of variance in attitudes about the product, F (4, 162) =
34.93, p <. 001, adj.R2 = .45. Among the first-order predictors, vividness of statistical
review was not significantly related to attitudes about the product, B = .07, t = 1.14., p =
.26, sr = .07, and vividness of narrative review was not significant, B = -.07, t = -1.16, p =
.25, sr = -.07. On the other hand, valence of statistical review was a significant predictor,
B = .37, t = 6.33, p < .001, sr = .36, indicating that attitudes about the product were more
positive in the statistical positive review condition than in the statistical negative review
condition. Valence of narrative review was a significant predictor, B = .55, t = 9.07, p <
.001, sr = .52, indicating that attitudes about the product were more positive in the
narrative positive review condition than in the narrative negative review condition.

The second-order predictors in the second block of the regression model
explained a significant amount of variance in attitudes about the product, F change (3, 159)
= 4.33, p = .006, R2 change = .04. The interaction of statistical review valence by statistical
review vividness was not significant, B = -0.25, t = -1 .51, p = .13, sr = -.08. The
interaction of statistical review valence by narrative review valence was not significant, B
= -0.62, t = -1.81, p = .07, sr = -.10. However, the interaction of narrative review valence
by narrative review vividness was significant, B = 0.47, t = 2.55, p = .01, sr = .14. Simple
slope of narrative review vividness was negative for narrative negative review, b = -O.3 7,
p = .02, whereas simple slope of narrative review vividness was positive but not
significant for narrative positive review, b = 0.10, p = .42. This finding indicated that for

narrative negative review, the more vivid the review was, the less positive attitudes were.

32

Intention to Purchase the Product

The overall model with all predictors was significant, F (7, 159) = 10.10, p <
.001, adj.R2 = .28. The first-order predictor in the first block of the regression model
accounted for a significant amount of variance in intention to purchase the product, F (4,
162) = 14.84, p <. 001, adj.R2 = .25. Among the first-order predictors, vividness of
statistical review was not significant, B = .08, t = 1.1 8, p = .24, sr == .08, and vividness of
narrative review was not significant, B = -.03, t = -0.46, p = .64, sr = -.03. On the other
hand, valence of statistical review was a significant predictor, B = .25, t = 3.73, p < .001,
sr = .25, indicating that intention to purchase the product was stronger in the statistical
positive review condition than in the statistical negative review condition. Valence of
narrative review was a significant predictor, B = .45, t = 6.28, p < .001, sr = .42,
indicating that intention to purchase the product was stronger in the narrative positive
review condition than in the narrative negative review condition.

The second-order predictors in the second block of the regression model
explained a significant amount of variance in intention to purchase the product, F change
(3, 159) = 3.04, p = .03, R2 change = .04. The interaction of statistical review valence by
statistical review vividness was not significant, B = 0.05, t = 0.25, p = .80, sr = .02. The
interaction of statistical review valence by narrative review valence was not significant, B
= -0.47, t = -1.25, p = .21 , sr = -.08. However, the interaction of narrative review valence
by narrative review vividness was significant, B = 0.51, t = 2.52, p = .01, sr = .16. Simple
slope of narrative review vividness was negative for narrative negative review, b = -0.33,
p = .03, whereas simple slope of narrative review vividness was positive but not

significant for narrative positive review, b = 0.16, p = .26. This finding indicated that for

33

narrative negative review, the more vivid the review was, the weaker intention to

purchase the product was.

34

Summary of Study 1 and Study 2 Findings

The two main hypotheses of the current study predicted stronger persuasive
effects of statistical review than narrative review and expected the difference between
statistical and narrative reviews to be greater when the reviews were negative than when
they were positive. Individuals considered statistical review more credible than narrative
review. Negative review was perceived as more credible than positive review. Statistical
negative review was perceived as more credible than narrative negative review, while
credibility of statistical positive review did not differ fi'om that of narrative positive
review. The data were consistent with hypotheses 1 and 2 only for review credibility, but
not for attitudes about the product and intention to purchase the product. Statistical
review did not differ fiom narrative review in affecting attitudes about the product or
intention to purchase the product. Compared to the control condition where individuals
did not read any reviews, the extent to which negative review affected attitudes (in terms
of decreasing positive attitudes) was greater than the extent to which positive reviews
affected attitudes (in terms of increasing positive attitudes). But the extent to which
negative reviews weakened intention to purchase the product did not differ from the
extent to which positive review strengthened intention. For negative review, statistical
review and narrative review did not differ in affecting attitudes about the product and
intention to purchase the product. Similarly, for positive review, statistical review and
narrative review did not differ in affecting attitudes about the product and intention to
purchase the product.

Study 1 of current research found that individuals Considered narrative reviews to

be more vivid than statistical reviews and that vividness was a moderator for the effect of

35

review type and valence on credibility perception of review and intention to purchase the
product. The more likely individuals were to perceive statistical negative review to be
vivid, the less likely individuals were to perceive it to be credible, whereas the more
likely individuals were to perceive statistical positive review to be vivid, the more likely
individuals were to perceive it to be credible. On the other hand, the more likely
individuals were to perceive narrative negative review to be vivid, the more likely
individuals were to perceive it to be credible, whereas the more likely individuals were to
perceive narrative positive review to be vivid, the less likely individuals were to perceive
it to be credible. When it comes to purchasing intention, the more likely individuals were
to perceive the statistical review to be vivid, the stronger were their intentions to purchase
the product. On the other hand, the more likely individuals were to perceive the narrative
review to be vivid, the weaker were their intentions to purchase the product.
Additionally, study 2 of current research found that when statistical negative
review was accompanied by narrative negative review, the reviews were perceived as the
most credible, compared to any other combination of statistical positive review and
narrative negative or positive review. The interaction of statistical review vividness with
statistical review valence was significant for review credibility, whereas the interaction of
narrative review vividness with narrative review valence was significant for attitudes
about the product and intention to purchase the product. That is, for statistical negative
review, the more likely individuals were to perceive the review to be vivid, the less likely
individuals were to perceive it to be credible. For narrative negative review, the more

likely individuals were to perceive the review to be vivid, the less positive individuals'

36

attitudes were about the product and the less likely individuals were to intend to purchase

the product.

37

DISCUSSION

The current research examined persuasive effects of online reviews for the
categories of statistical versus narrative and positive versus negative reviews. Past
research showed that the usefulness of negative reviews varied with product types (Sen &
Lerman, 2007) and consumers' familiarity with a retailer (Chatterjee, 2001). The current
study adds review type (summary of numeric ratings as a type of statistical evidence
versus consumer comments as a type of narrative evidence) to the literature as another
moderator of the effects of negative reviews on credibility perception of reviews.
Statistical negative reviews were perceived as more credible than narrative negative
reviews, possibly because a large number of ratings indicates greater objectivity, less
bias, and numerous people who were familiar with the product. Sen and Lerman (2007)
and Chatterjee (2001) argued that the stronger effects of negative reviews for utilitarian
products and unfamiliar retailers could be because negative reviews for utilitarian
products had greater objectivity and negative reviews of unfamiliar retailers helped with
reducing uncertainties about risks. These reasons can be relevant for the stronger
credibility of negative reviews containing statistical information, because statistical
reviews may have heightened perceptions of objectivity and amount of risk information
provided by experienced users of the product.

Because previous research findings have shown greater persuasiveness of
negative WOM over positive WOM (Herr et al., 1991; Larczniak et al., 2001), it was not
surprising for the current study to find that negative reviews were perceived as more
credible than positive review, that statistical negative reviews coupled with narrative

negative reviews were the most credible, and that negative reviews affected attitudes

38

about the product to a greater extent than positive reviews did. Nevertheless, a reason for
the current finding could be that a utilitarian product (a digital camera) was used in the
current study. Sen and Lerman (2007) showed that consumers were more likely to
consider negative reviews useful for utilitarian products (e.g., cell phones and digital
cameras) than for hedonic products (e.g., movies and music CDs and DVDs), possibly
because reviews for utilitarian products could be more objective and based on actual
features and technical functions of the products, whereas reviews for hedonic products
could be more subjective and based on individuals' personal preferences. It is possible
that by considering negative reviews more credible than positive reviews, participants in
the current study might have considered the negative reviews to be genuine concerns and
sincere warnings expressed by previous buyers who hope others do not experience the
same unpleasantness and disappointment. On the other hand, participants in the current
study might have considered the positive reviews to be consiStent with the usual product
information as advertised by manufacturers.

Another reason for the current finding could be that participants did not know the
company that made and sold the digital camera used in the current study (i.e., the digital
camera was labeled as "Brand X Camera" with no specific information about who made
the camera). Chatterjee (2001) reported that consumers who chose a retailer unfamiliar to
them tended to look at more negative reviews than those who chose a familiar retailer,
possibly because negative reviews could be more helpful for consumers to avoid risks
with an unfamiliar retailer. Chatterjee (2001) firrther speculated that users of a familiar
retailer could be more likely to dismiss negative reviews because they were already

satisfied with the retailer. Thus, a future study may examine if the effects of negative

39

reviews on credibility perception vary with people's familiarity with manufacturers and
sellers.

Unlike the difference between statistical and narrative reviews for review
credibility, the current study did not find many differences between statistical and
narrative reviews for attitudes about the product and intention to purchase the product.
One possible explanation could be that attitudes about the product and intention to
purchase the product could be influenced by many factors other than the reviews which
participants had just read. Naturally, positive reviews improved attitudes about the
product and increased intention to purchase the product, while negative reviews
decreased them. But attitudes about the product and purchase intention could have
depended more on economic status, current necessity of a new product, and opportunities
to compare various Web sites, reviews, and other similar products, rather than on how
many others rated the product versus how others described their experience with the
product.

The current study measured individuals' evaluations of how vividly each review
type was presented, predicting that vividness would explain differences between narrative
and statistical reviews for their effects on credibility, attitudes about the product, and
intention to purchase the product. Although narrative reviews were perceived as more
vivid than statistical reviews, it was statistical reviews that were perceived as more
credible regardless of vividness. The current findings indicated that vividness did not
explain the persuasive effects of narrative reviews, and that vividness did not have direct
effects on review credibility, attitudes about a product, and intention to purchase the

product.

40

The role of vividness varied across different review types. Vividness of statistical
negative review was inversely related to review credibility, but vividness of statistical
positive review was positively related to review credibility. Vividness of narrative
negative review was positively related to review credibility, but vividness of narrative
positive review was negatively related to review credibility. One possible explanation
could be that individuals might have interpreted vividness of statistical negative review to
be unreal (e. g., how can this many people react negatively about a product?, how can a
product with this many negative review be still advertised for sale?) On the other hand,
individuals might have interpreted vividness of narrative negative review as a serious
expression of consumers' concerns about the product, whereas vividness of narrative
positive review might have led individuals to question the real intention of the review
comments. Sen and Lerman (2007) demonstrated that people were more likely to
attribute product related motivations to negative review of a utilitarian product than that
of a hedonic product. But their study did not differentiate numeric and narrative reviews
for motive attribution. The current study findings may imply that people may make
different attributions about negative and positive reviews when the reviews are in the
forms of numeric ratings versus narrative comments.

Another implication of the current findings regarding vividness is that individuals
varied in their perceptions of the vividness of the same review. Such individual variations
in vividness were differentially related to review credibility, attitudes about the product,
and purchase intention across different types and valences of reviews. For example, the
current study found that vividness of statistical review was positively related to

purchasing intentions, but vividness of narrative review was negatively related to

41

purchasing intentions. Vividness of statistical negative review was inversely related to
review credibility, but vividness of narrative negative review was inversely related to
attitudes about the product and also inversely related to purchasing intentions. Because
the current study did not examine what caused such individual variations of vividness of
online reviews and subsequent variations in review credibility, attitudes about the
product, and purchase intention, it is not easy to explain the findings concerning
vividness. One speculation, however, may point to the trend that improvement in Web
technologies enables online reviews to be presented with colorful graphics and/or in
dynamic forms. How graphically and vividly Web sites present online reviews can vary
as well. Individuals' experiences with various Web sites might have led to comparisons of
the current study manipulation materials against the review formats of Web sites
individuals were already familiar with. Thus, the current research materials might have
seemed less vivid than usual for some participants and more vivid for others.
Limitations of Current Research and Suggestions for Future Research

A few limitations of the current study may provide suggestions for future
research. First, only one product (i.e., a digital camera) was used in the current study. As
previous research showed, product type can moderate the effects of eWOM (e.g., Sen &
Lerman, 2007). Furthermore, the current study focused only on product review and not
on retailer or seller review concerning shipping and handling quality, return policy, and
customer service quality. Future research can examine whether the current findings can
be generalized to online reviews of other types of product such as CDs and DVDs of

music and movies and reviews of retailers or sellers.

42

The second limitation may pertain to the number of reviews shown in the
statistical review versus the narrative review. Although the current study tried to prepare
the equal base rate (i.e., the total number of reviews) in the statistical and narrative
review conditions, participants in the current study might have not noticed it. For the
statistical review, participants could easily notice that the rating summary was based on
1000 feedback ratings. For the narrative review, although a phrase, "1000 feedback
received," was included at the top of 15 narrative review comments, the phrase might
have been in too small of a font size for participants to notice. Thus, at this point, it is
unclear whether participants in the narrative review condition counted or discounted the
base rate when evaluating the review. Additionally, it can be asked if the number (15) of
narrative review comments might have affected the current findings. Future research may
examine if the varying numbers (e.g., 15 versus 30) of narrative review comments shown
to participants make a difference and if there is a threshold of narrative review comments
that consumers prefer to read before making a purchase decision.

Third, the current study did not provide opportunities for people to choose which
reviews they want to check. When the amount of reviews and feedbacks available is too
large to check all, online consumers may chose to read only several review comments
and their choices can be guided by various factors. Consumers may choose to read longer
reviews because longer reviews are often perceived as more helpful (Chevalier &
Mayzlin, 2006; Sen, 2008). Some consumers may prefer the reviews by someone with
good reputation. For example, some Web sites provide a way for people to rate the
review itself and some consumers may choose to read a review comment that many

people already find usefirl.

43

CONCLUSION

Advances in technologies and increases in online communication result in
growing amounts and diversifying qualities of information available to consumers.
Understanding how marketers and consumers adapt to any new modes of online
communication and how they process various forms of online information can remain
constant challenges as the speed of technology improvement increases. The current study
findings can have some implications for scientific research, commercial practice, and
uses of online reviews as a form of eWOM. Careful management of negative reviews can
be a significant way to influence attitudes and intention to purchase about a given
product. Effective control and presentation of numeric ratings as statistical reviews can
be important ways to increase review credibility and attract consumers to check the

reviews on a given Web site.

Table 1 Demographic Information (Study 1)

 

 

 

Category Number of Subject Percentage
Gender (Age M = Male 57 31.5
22.35, SD = 3.09)
Female 124 68.5
Grade Sophomore 4 2.2
Junior 69 38.5
Senior 51 28.5
Graduate 55 30.7
Ethnicity Caucasian/European 1 23 69.5
American
Hispanic 1 0.6
Afiican American .8 4.5
Pacific Islander 4 2.3
Asian American 15 8.5
Mixed 1 0.6
Other 25 14.1
* n = 179

45

Table 2 Online shopping Tendency (Study 1)

 

 

M Median SD
Average spending 59.83 30 112.11
(per month) (U. S. dollar) (U. S. dollar)
Average purchase 2.27 1 5.86
(per month) (times) (times)
To what extent you believe online 5.45 6 0.97
review is useful
(1 = lowest and 7 = highest)
To what extend you believe it is useful 4.96 5 1.36

to rely on online reviews for
purchasing decisions

(1 = lowest and 7 = highest),

 

46

Table 3 Demographic Information (Study 2)

 

 

 

Category Number of Subject Percentage
Gender (Age M = Male 55 32.5
21.48, SD = 2.07)
Female 114 67.5
Grade Freshman 2 1.2
Sophomore 37 21.8
Junior 50 29.4
Senior 73 42.9
Graduate 8 4.7
Ethnicity Caucasian/European 1 30 76.9
American
Hispanic 2 1.2
Afiican American 6 3.6
Asian American 14 8.3
Mixed 4 2.4
Other 13 7.7
* n = 170

47

Table 4 Online shopping Tendency (Study 2)

 

 

M Median SD
Average spending 45.77 25 53.42
(per month) (U. S. dollar) (U. S. dollar)
Average purchase 1.43 0.83 2.28
(per month) (times) (times)
To what extent you believe online 5.49 6 0.93
review is useful
(1 = lowest and 7 = highest)
To what extend you believe it is useful 5.01 5 1.22

to rely on online reviews for
purchasing decisions

(1 = lowest and 7 = highest)

 

48

 

Table 5 Comparisons of Experimental Conditions in Study 1 and Study 2

 

 

R . Attitudes Intention to
Cid?” about a Purchase a
Credrbrlrty” ProductM Product”
Statistical Negative Review
coupled with Narrative Negative 39 5.16q (1.24) 2.15x (0.98) 1.933 (0.86)
Review
Statistical Negative Review only 36 5.1 1q (1.20) 2.16K (1.09) 1.96a (0.96)
Narrative Negative Review only 34 4.43r (0.72) 2.53K (0.90) 217,002)
Statistical Positive Review coupled
with Narrative Negative Review 45 4.331r (0.78) 3.69y (1.34) 3.031, (1.40)
. . 3.20bc
Control group — no revrew included 36 -- 4.16y (0.76) (1 07)
Statistical Negative Review 3 47
coupled with Narrative Positive 42 4.24r (0.73) 419,025) (1' 3;;
Review '
Statistical Positive Review coupled 3.89“)
with Narrative Positive Review 44 4'37’ (0'75) 49% (0'79) (1.09)
Statistical Positive Review only 39 4.48r (0.65) 5.00Z (0.82) 4.49,, (1.1 1)
Narrative Positive Review only 36 4.17,(0.80) 5.062 (0.64) 4.30,, (0.91)

 

Note. Standard deviations are in parentheses.

_ Means not sharing a same subscript within each column differ from one another at p < .05.
”ANOVA with credibility as the dependent variable, F (7, 307) = 7.24, p < .001, n2 = .14.
”ANOVA with attitudes as the dependent variable, F (8, 342) = 57.08, p < .001, n2 r .57.

”ANOVA with intention as the dependent variable, F (8, 342) = 28.96, p < .001, n2

49

II'

.40.

 

APPENDIX A
Background Information

Plege answer the following questions about yourselfi

B1. Have you ever seen online reviews? (for example, feedback from Amazon or eBay)

El Yes [:1 No
B2. Have you ever bought any product online?
Cl Yes D No

B2a. If yes, how often do you shop online?
times per D week El month Cl year

 

B2b. If yes, how much do you spend while shopping online?
$ per El week El month El year

 

Please indicate the extent to which you agree or disagree with each sentence below.

B3. I use online reviews to decide what I purchase with online shopping.

 

 

 

 

 

 

 

 

 

 

 

 

Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
B4. I use online reviews to decide what I purchase with offline shoppinL
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7

 

 

 

 

 

 

 

 

B5. I prefer the website with online reviews rather than the website without online

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

reviews.
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
B6. I often check online reviews to get information about products.
Strongly . Somewhat Somewhat Strongly
Disagree Drsagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
B7. I have to check online reviews before any purchase is made. _
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
B8. I usually check online reviews before finalizin my purchase.
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7

 

 

 

 

 

 

 

 

50

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

B9. I believe it is useful to rely on online reviews for purchasin r, decisions.
Strong] . Somewhat Some hat Strongly
Disagre: Disagree Disagree Neutral Agr‘Ze Agree Agree
1 2 3 4 5 6 7
BIO. I believe the information from online reviews are usefirl.
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
Please answer the following Questions about yourself
D1. Age
D2. Gender E] Male [:1 Female
D3. Major
D4. Freshman Sophomore Junior Senior Graduate
D5. Please indicate your ethnicity (check one)
Caucasian/European American Hispanic
Afiican American Pacific Islander
Native American Mixed (please specify
Asian American Other (please specify

Please go on to the next page

 

51

 

 

 

 

 

 

35
mm. oém .. i of... . .0 Erefiggaggimssm “Rama
F o o 32 9 s; acreage;
:5. ,, sex SEE—«Susana:
.29on .3: Em 52% 38
com com 8 _. o m: So 33 ”as 33 .810 22E Em
“£86 2 2:36 o 535 L 8.33 23..
MM. ESE 2 new .— 3350“ 35233333 ”$53.23?
B 9 Egg 325.3 83 v5.29: :3. :5 . Ema.— .A “23%“;
m .9 Ease Eaa 2: ass a 2 2 ea . 5
mm. s e 32% 3
m 2 3 £03 5 . 53 .3 2:.
x
M 2
$8 ham fiwnmomm .gEmom 3533.5 assuage
senses ............ 33.2; 383
m 3.33833 33;
m
E . H. .
.m 3.2m was: .3535
m a 5%
e . . . , c
“m. @533 5:: =2... .Efihm 53m ECU «QWEE
o . . :1:
P

52

Narrative Statistic Evidence

.8 g. .2 ,. ea;
. o o .2329

g s 2 as...
3.2... N. 2.5... o 58.... .
..,,~. .mfieoEm 1.2... H gogom .

.39.... 3.5... 82 as 22.
.. see a... a... so...

avg—coo. 2.. m. 23 :23 2:

go. .23. “8.33 3.8..

3.533533. 3...... -

e... w is. as“. 55.52 2 u. o> .

E a 3.335.

E... ”sausage...
E802... Eggs 3......
85.3. as

.228 355.5 5...... “3.3. 32%?

 

3.... 3..
.5 3 x .8 2 x .5 z. ”.593. 33......
as...

 

3E“ 5.... a. 3 NE 3.2.. 5.3 a $2.5

 

wen—hug” “ESE—E so:

2&3: F. m at, 22.8 x23

3a3§fiaefis 331...... 5.2.3 .v

 

1, $55 Wages .3333

 

a
3...... Eu... 3 .5 Luv. E

 

 

_. Em

Eou§§6

53

Positive Narrative Evidence

REViEW 2 The below is the part of feedback about the digital camera x

from the previous buyers.
1” mm (viewing 1-25)

Mm
0 Alazing quality.

0 This is not that heavy so that I can carry easily.
0 This is the one I an looking for.
0 Beautiful design!

0 It’s acceptable.
6Wenlusedthisatthebirthdayparty ‘ "
everyone loved their face in the calera.

9 Nice one. I recouend this to my pals.

 

...~ --.......... -..-.. M- W.....

0 A11 functions are fantastic!

. ....__..._..-..... .................. mmw-

0 This was ly anniversary gift and Iy husband loved it. It works well.

ONOTBAD

0 I an pleased with the quality of the picture...

0 Useful product. I an happy with oy purchase.

0 This canera changed ny weekend in good way

—-- -wflw. _. ._.. .4... .. -.. - fi- .

0 I will race-Tend this to everyone.
0 Nice appearance, this is what I wanted

54

From Buyer
Buyer. whitefiioufi)

Buyer. pmhalekw(203 *)

Buyet: mor_chick87 (39 {k )

Buyer tennpwzons {in

Buyer. siar¢20_3 ( 24 fl!)

.........‘.......1.;"'
"a...".....,4... ' '

"...J,...... 9.5 ,'

“.....‘............5

Buyer: momma *)

Buyer ruppel‘1969(104 fir )
Buyer. pdfner309 ( 3)

Buyer; oolfer110751( 104 fi )

Buyer. 25101151 * ,1

Buyer. luypmmytibg * )

 

Negative Narrative Evidence

Review 2 The below is the part of feedback about the digital calera x

fron the previous buyers.
1mhdndmamd (vicvang 1-25)
Feedioct

0 Disappointing quality.
{thie'i‘ssd heavy d... c... hottahyeasiiy.
*5 hi; i... the mama... e...

G hwful‘ design!

‘ at}; acceptable.
"5h... I used thisat the birthday party. ’ ’

no one liked their face in the cawera.

° Bad one. I can’t recomend this to ly pals.

0 A11 functions are horrible!

O This was my anniversary gift, but my husband hated it.It works badly.

. -.-.-.... 7.....1 ..._. th..-

quorum

0 I an not pleased with the quality of the picture-

0 Not useful product. I regret wy purchase.

9 This cawera ruined my weekend

0 I won’t recomend this to anyone.

9 Disappointing appearance, this is not what I wanted.

55

From Buyer

Buyer caniesm (6)
Buyer jimhaieitw( 2113 fir ) 4 1
Buyer. hogmmme :3; )
Buyer kennpttlz‘oi 36 fi)
a... smzo_3(22.i 3%)

Buyer aceohanshteuoms m

Buyer gsharkytS)

Buyer. 7442.38171120 {k >

Buyer. susanandrandygs ( 50 *)

.u.....hst..g,‘ '

B‘i‘lei NW911969(1104*)‘
3W9" parlance ( 3 J
Buyer. codefl10751 . 104 it }
Buyer. 25m 151 *,

Buyer lilgimmfl 169 fir )

APPENDIX C

Vividness / Credibility / Attitudes about a Product / Purchase Intention

Afler reading Review 1.

Plea_se indicate the extent to which you agree or disagree with each statement below.

MSl. Overall reviews are positive.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
M82. Almost all of reviews are recommending this product.
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
M83. People who contribute reviews are mostly satisfied with this product.
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
M84. Overall reviews are negative.
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
M85. Almost all of reviews are not recommending this product.
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
M86. People who contribute reviews are mostly unsatisfied with this product
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
M87. It seems that people generally recommend this product.
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
M88. It seems that people do not generally recommend this product.
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7

 

 

 

 

 

 

 

 

56

 

 

 

 

 

 

 

 

 

Please indicate your feeling about Review 1 with each stgtement below.

V68 1. In general, to what extent were these messages vividly presented?
I Not at all 1 2 3 4 5 6—7 Very much ]

 

 

 

 

 

 

VPSl. To you, personally, to what extent were these messages vividly presented?
[ Not at all 1 2 3 4 5 5. 7 Very much j

 

 

 

 

 

VG82. In general, to what extent were these messages emotionally presented?
[ Not at all 1 2 3 4 5——-6————7 Very much 1

 

 

 

 

 

VPSZ. To you, personally, to what extent were these messages emotionally presented?
[ Not at all 1 2 3 4 5——6—-—7 Very much

 

 

 

 

 

VGS3. In general, to what extent were these messages concretely presented?
[ Not at all 1 2 3 4 5—6—7 Very much ]

 

 

 

 

 

VPS3. To you, personally, to what extent were these messages concretely presented?
[ Not at all 1 2 3 4 5 6 7 Very much I

 

 

 

VGS4. In general, to what extent were these messages imaginably presented?
L Not at all 1 2 3 4 5 6———7 Very much I

 

 

 

 

 

 

VPS4. To you, personally, to what extent were these messages imaginably presented?
F Not at all 1 2 3——-4—--—5—--—6——7 Very much

 

 

 

 

P81. How persuasive do you think these messages were?
I Not at all 1 2 3 4 5 5 7 Very much |

 

 

 

 

P82. To what extent do you think your opinions on this product were influenced
by these messages?
I Not at all 1

 

2 3 4 s 5 7 Very much ]

 

 

 

 

57

After reading Review 2,

Please indicate the extent to which you agree or disagree with each statement below.

MNl . Overall reviews are positive.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Strongly . Somewhat Somewhat Strongly
Disagree Dlsagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
MNZ. Almost all of reviews are recommending this product.
Strongly . Somewhat Somewhat Strongly
Disagree Dlsagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
MN3. People who contribute reviews are mostly satisfied with this product.
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
MN4. Overall reviews are negative.
Strongly . Somewhat Somewhat Strongly
Disagree Dlsagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
MNS. Almost all of reviews are not recommending this product.
Strongly . Somewhat Somewhat Strongly
Disagree Dlsagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
MN6. People who contribute reviews are mostly unsatisfied with this product
Strongly . Somewhat Somewhat Strongly
Disa Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
MN7. It seems that people generally recommend this product.
Strongly . Somewhat Somewhat Strongly
Disagree Disagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
MN8. It seems that people do not generally recommend this product.
Strongly . Somewhat Somewhat Strongly
Disagree Dlsagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7

 

 

 

 

 

 

 

 

58

 

 

 

 

 

 

 

 

 

Please indicate your feeling gbout Review g with each statement below

VGNl. In general, to what extent were these messages vividly presented?
I Not at all 1 2 3 4 5———-6——7 Very much I

 

 

 

 

 

VPNl. To you, personally, to what extent were these messages vividly presented?
I Not at all 1 2 3 4 5 6—-—-—7 Very much I

 

 

 

 

VGN2. In general, to what extent were these messages emotionally presented?
I Not at all 1 2 3 4 5———-6—-——-7 Very much I

 

 

 

 

VPN2. To you, personally, to what extent were these messages emotionally presented?
I Not at all 1 2 3 4 5 6—7 Very much I

 

 

 

 

 

 

VGN3. In general, to what extent were these messages concretely presented?
I Not at all 1 2 3 4 5 6 7 Very much I

 

 

 

 

 

VPN3. To you, personally, to what extent were these messages concretely presented?
F Not at all 1 2 3 4 5 6—7 Very much I

 

 

 

 

VGN4. In general, to what extent were these messages imaginably presented?
I Not at all 1 2 3 4 5 6———-—7 Very much I

 

 

 

 

 

VPN4. To you, personally, to what extent were these messages imaginably presented?
I Not at all 1 2 3 4 5—-—6———7 Very much

 

 

 

 

 

PNl. How persuasive do you think these messages were?
I Not at all 1 2 3 4 5 5——-—7 Very much I

 

 

 

 

 

PN2. To what extent do you think your opinions on this product were influenced by
these messages?
I Not at all 1 2 3 4 5

 

-6———-—-—7 Very much I

 

 

 

59

After reading both Review 1 and Review 2, what do you think about these reviews?

Cl. To me, overall reviews are,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Extremely Very Somewhat Neutral Somewhat Very Extremely
Biased Biased Biased Unbiased Unbiased Unbiased
l 2 3 4 5 6 7
C2. To me, overall reviews are,
Extremely Very Somewhat Neutral Somewhat Very Extremely
Trustworthy Trustworthy Trustworthy Untrustworthy Untrustworthy Untrustworthy
l 2 3 4 5 6 7
C3. To me, overall reviews are,
Extremely Very Somewhat Neutral Somewhat Very Extremely
Inaccurate Inaccurate Inaccurate Accurate Accurate Accurate
l 2 3 4 5 6 7
C4. To me, overall reviews are,
Extremely Very Somewhat Neutral Somewhat Very Extremely
Unbelievable Unbelievable Unbelievable Believable Believable Believable
l 2 3 4 5 6 7
C5. To me, overall reviews are,
Extremely Very Somewhat Neutral Somewhat Very Extremely
Incomplete Incomplete Incomplete Complete Complete Complete
1 2 3 4 5 6 7

 

 

 

 

 

 

 

 

6O

 

 

 

 

 

What do you think about this product? Please imag'ne you need to buy a new digital
camera, then indicate your preference on the 71point scale.

A1. Please indicate your feelings about this product, if it is on the market.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Extremely Very Somewhat Neutral Somewhat Very Extremely
Bad Bad Bad Good Good Good
1 2 3 4 5 6 7
A2. Please indicate your feelings about this product, if it is on the market.
Extremely Very Somewhat Neutral Somewhat Very Extremely
Unfavorable Unfavorable Unfavorable Favorable Favorable Favorable
l 2 3 4 5 6 . 7
A3. Please indicate your feelings about this product, if it is on the market.
Extremely Very Somewhat Neutral Somewhat Very Extremely
Undesirable Undesirable Undesirable Desirable Desirable Desirable
l 2 3 4 5 6 7

 

 

 

 

 

 

 

 

61

 

 

 

 

Please imagine thatjou may be looking to buy a digital camera in the near future and
then indicate the extent to which you agree or disagree with each sentence below.

Pleaie circle a number that best reflects your agreement level.

BIl. If I am going to buy a new digital camera in the future, I intend to buy this product

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Strongly . Somewhat Somewhat Strongly
Disagree Dlsagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
BIZ. If I am going to buy a new digital camera in the future, I have no plan to buy this
product
Strongly . Somewhat Somewhat Strongly
Disagree Dlsagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7
B13. If I am going to buy a new digital camera in the future, I have it in mind to buy this
product
Strongly . Somewhat Somewhat Strongly
Disagree Dlsagree Disagree Neutral Agree Agree Agree
1 2 3 4 5 6 7

 

 

 

 

 

 

 

 

B14. If I am going to buy a new digital camera in the future, I mean to buy this product

 

 

 

 

 

 

 

 

 

Strongly Disagree Somewhat Neutral Somewhat Agree Strongly
Dlsagree Dlsagree Agree Agree
1 2 3 4 5 6 7

 

BIS. If I am going to buy a new digital camera in the filture, I would not buy this product

 

 

 

 

 

 

 

 

 

Stron l . Somewhat Somewhat Stron I
Disagfie: Dlsagree Disagree Neutral Agree Agree Agrcfey
1 2 3 4 5 6 7

 

Thank you for participating in this study

62

 

 

 

 

 

 

REFERENCES

Abernathy, E. (1964) .The advocate: A manual of persuasion. New York: Mckay.

Allen, M., & Preiss, R. W. (1993). Replication and meta-analysis: A necessary
connection. Journal of Social Behavior and Personality, 8, 9-20.

Allen, M., & Preiss, R. W. (1997). Comparing the persuasiveness of narrative and
statistical evidence using meta-analysis, Communication Research Report, 14,
125-1 3 1 .

Allsop, D. T. Bassett, B. R., & Hoskins, J. A. (2007). Word-of mouth research: Principles
and applications, Journal of Advertising Research, 47, 398-411.

Amdt, J. (1967a). Role of product-related conversations in the diffusion of a new product,
Journal of Marketing Research, 4, 291 -295.

 

Amdt, J. (1967b). Word-of mouth advertising: A review of the literature. New York:
Advertising Research Foundation.

Ameson, P. A., & Query, J. L. (2001) .The case study: Revitalizing a nonprofit health
organization. In 8. L. Kreps (Eds), Qualitative research: Applications in
organizational life (2nd ed.) (pp. 151-169) .Cresskill, NJ: Hampton Press.

Baesler, E. J ., & Burgoon, J. (1994). The temporal effects of story and statistical evidence
on belief change. Communication Research, 21, 582-602.

Bansal, H. 8., & Voyer, P. A. (2000). Word of mouth processes within a service purchase
decision context. Journal of Service Research, 3, 166-177.

Bass, F. M. (2004). A new product growth model for consumer durables. Management of
Science. 50, 1825-1832

Beal, G. M., & Rogers, E. M. (1957) Informational success in the adoption process of
new fabrics. Journal of Home Economics, XLIX, 630-634.

Bickart, B., & Schindler, R. M.(2001). Internet forunls as influential sources of
consumer information, Journal of Interactive Marketing, 15, 3, 31-40.

Biyalogorsky, E., Gersyner, E., & Libai, B. (2001). Customer referral management:
Optimal reward programs. Marketing Science, 20, 82-95.

Borgida, E., & Nisbett, R. E (1977). The differential impact of abstract vs. concrete
information on decisions. Journal of applied Social Psychology, 7, 258-271.

63

Boster, R, Cameron, K. A., Campo, 8., Liu, W.-Y., Lillie, J., Baker, R., & Yun, K. A.
The persuasive effects of statistical evidence in the presence of examplars.
Communication Studies, 51, 296-306.

Brooks, R. C. Jr. (1957) .Word-of mouth advertising in selling new product. Journal of
Marketing, 22, 154-161.

Brown, J. J ., & Reingen, P. H. (1987). Social ties and word-of-mouth referral behavior.
Journal of Consumer Research, 14, 350-362.

Bureau of Applied Social Research, Colombia University, On the Flow of Scientific
Information. In the Medical Profession (2 Vols.), unpublished research reports,
1955. (James Coleman, Elihu Katz, and Herbert Menzel published in Book form)

Cacioppo, J. T., Petty, R. E., Kao, C. F., & Rodriguez, R. R. (1986). Central and
peripheral routes to persuasion: An individual difference perspective. Journal of
personality and Social Psychology, 51(5), 1032-1043.

 

Chevalier, J. A., & Mayzlin, D. (2006). The effect of word of mouth online: Online book i
reviews, Journal of Marketing Research, 43, 345-3 54.

Charlett, D., & Garland, R. (1995), How damaging is negative word of mouth?
Marketing Bulletin, 6, 42-51.

Chatterjee, P. (2001). Online reviews-do consumers use them? Advances in Consumer
Research, 28, 129-133.

Chen, Y., & Xie, J. (2008). Word-of-mouth as a new element of marketing
communication mix, Management Science, 54, 477- 491 .

Cohen, M. R., & Nagel, E. (1962). An introduction to logic. New York: Harcourt, Brace
&World.

Czepiel, J. A. (1974). Word-of-mouth process in the diffusion of a major technological
innovation. Journal of Marketing research, 1 I , 172-180.

Davis, D., Guiltman, J. P., 82 Jones, W. H. (1970). Service characteristics, consumer
search and the classification of retailing services. Journal of Retailing, 55(3), 3-23.

Dewan, 8., & Hsu, V. (2004). Adverse selection in electronic market: Evidence from
online stamp auctions. Journal of Industrial Economics, 52, 497-516.

Dellacoras, C. (2003). The digitization of word-of-mouth: Promise and challenges of
online feedback mechanisms. Management Science, 49, 1401 -1424.

64

Dickson, P. R. (1982). The impact of enriching case and statistical information on
consumer judgments. Journal of Consumer Research, 8, 398-406.

Ditcher, E. (1966). How word-of-mouth advertising works. Harvard Business Review, 44,
147-166.

Dwyer, P. (2007). Measuring the value of electronic word of mouth and its impact in
consumer communities. Journal of Interactive Marketing, 21 (2), 63-79.

Ellinson, L. L., & Buzzanell, P. M. (1999). Listening to women’s narratives of breast
cancer treatment: A feminist approach to patient satisfaction with physician-
patient communication. Health Communication, 1 I, 153-183.

Evans, M., Wedande, G., Ralston, L., & Van t Hul, S. (2001). Consumer interaction in
the virtual era: some qualitative insights, Qualitative Market Research: An
International Journal, 4, 150-159.

Flanagin, A. J ., & Metzger, M. J. (2000). Perceptions of Internet information credibility.
Journalism and Mass Communication Quarterly, 77, 515-540.

Feldman, S. P. & Spencer, M. C. (1965). The Effect of Personal Influence in the
Selection of Consumer Services, in Bennett, P. (Eds), Marketing and Economic
Development, American Marketing Association, Chicago, IL, pp.440-52.

Fotheringham, W. C. (1966). Perspective on persuasion. Boston: Alyn and Bacon.

Gilly, M. C., Graham, J. L., Wolfinbarger, M. F ., & Yale, L. J. (1998). A dyadic study of
interpersonal information search. Journal of the Academy of Marketing Science,
26, 83-100.

Goldenberg, J ., Libai, B.,& Muller, E. (2001). Talk of the Network: A Complex Systems
Look at the Underlying Process of Word-of-Mouth, Marketing Letters, 12, 3,
21 1-223.

Greene, K., & Brinn, L. S. (2003). Message influencing college women’s tanning bed
use: Statistical versus narrative evidence format and a self-assessment to increase
perceived susceptibility. Journal of Health Communication, 8, 443-461.

Hamill, R., Wilson, T. D., & Nisbett, R. E. (1980). Insensitivity to sample bias:
Generalizing fiom atypical cases. Journal of Personality and Social Psychology,
39, 578-589.

Hennig-Thurau, T., Gwinner, K. P., Walsh, G., & Gremle, D. D. (2004) Electronic word-
of-mouth Via consumer-opinion platforms: What motivates consumers to
articulate themselves on the Internet?, Journal of Interactive Marketing, 18, 1, 38-
52.

65

Herr P. M., Kardes, F. R., & Kim, J. (1991). Effects of word-of-mouth and product
attribute information on persuasion: An accessibility-diagnosticity perspective.
Journal of Consumer Research, I 7, 454-462.

Heywood, K. M. (1989) Managing word of mouth communications, The Journal of
Service Research, 6, 287-303.

Hilton, D. J. (1995). The social context of reasoning: Conversational inference and
rational judgment. Psychological Bulletin, 118, 248-271.

Hoeken, H. (2001). Anecdotal, Statistical, and Causal Evidence: Their Perceived and
Actual Persuasiveness, Argumentation, 15, 4, 425-437.

Holmes, J. H., & Lett, J. D. (1977). Product sampling and word of mouth. Journal of
Advertising, 1 7, 35-40.

Hu, N., Liu, L., & Zhang, J. J. (2008). Do Online review affect product sales? The role of
reviewer characteristics and temporal effects. Information Technology
Management, 9, 201-214. E

 

Iyengar, 8., & Kinder, D. R. (1987). News that matters. Chicago: University of Chicago
Press.

Kalyanam, K., & McIntyre, 8. (2001) Return on reputation in online auction markets.
Santa Clara University, Santa Clara.

Kahneman, D., & Tversky, A. (1973). On the psychology of prediction. Psychological
Review, 80, 237-251.

Katz E. & Lazarfeld, P. F. (1995). Personal influence. Glencoe, 11: Free Press

Kazoleas, D. C. (1993). A comparison of the persuasive effectiveness of qualitative
versus quantitative evidence. A test of explanatory hypotheses. Communication
Quarterly, 41, 40-50.

Kiel, G. C., & Layton, R. A. (1981). Dimensions of consumer information seeking
behavior. Journal of Marketing Research, I 8, 233-239.

Koballa, T. R. (1986). Persuading teachers to reexamine the innovative elementary
science programs of yesterday: The effect of anecdotal versus data-summary
communications, Journal of Research in Science Teaching, 23, 5, 43 7-449.

Kopfrnan, J i E., Smith, 8. W., Yun, J. K., & Hodges, A. (1998). Affective and cognitive

reaction to narrative versus statistical evidence organ donation messages. Journal
of Applied Communication Research, 26, 279-300.

66

Lazarsfeld, P. F ., Berelson, B. R., & Gaudler, W.(1948). The people ’s choice. New York:
Columbia University Press.

Laczniak, R. N., DeCarlo, T. E., & Ramaswamy, S. N. (2001). Consumers’ responses to
negative WOM communication: An attribution theory perspective. Journal of
Consumer Psychology, 1 I, 57-73.

Liu, Y. (2006). Word of mouth for movies: Its dynamics and impact on box office
revenue. Journal of Marketing, 70, 74-89.

Martilla, J. A. (1971). Word-of-mouth communication in the industlial adoption process.
Journal of Marketing Research, 8, 173-178.

McCrosky, J. C. (1969). A summary of experimental research on the effects of evidence
in persuasive communication. Quarterly Journal of Speech, 55, 169-176.

McCrosky, J. C. (1967). The effects of evidence in persuasive communication. Western
Speech, 31, 189-199.

Mcdonald C. G., & Slawson, V. C. Jr. (2002). Reputation in an intemet auction market.
Economic Inquiry, 40, 533-650.

Melnik, M. 1., & Ahn, J. (2002). Does a seller’s reputation matter? Evidence fiom eBay
auctions. Journal of Industrial Economics, 50, 337-350.

Miller, G. (Victor D. Wall Jr.) (1972). Evidential attitudes and attitudes change. Western '
Speech, 36, 115-123.

Nadler, M. A. K. (1982). Evidence usage in persuasion. Dissertation Abstracts
International, 44, 910A.

Nisbett, R., & Ross, L. (1980). Human inference: Strategies and shortcomings of social
judgment. Englewood Cliffs, NJ: Prentice-Hall.

Nisbett, R. E., & Borgida, E. (1975). Attribution and the psychology of prediction.
Journal of Personality and Social Psychology, 32, 932-943.

O’Keefe, D. J. (2002). Persuasion: Theory and research (2nd ed.). Newbury, CA: Sage.

O’Keefe, D. J .(2004). Trends and prospects in persuasion theory and research. In J. S.
Seiter & R. H. Glass (Eds), Perspective on persuasion, social influence, and
compliance gaining ( 31-43). Boston, MA: Allyn & Bacon.

Park, C., & Lee, T. M. (2009). Information direction, website reputation and eWOM

effect: A moderating role of product type. Journal of Business Research, 62, 61-
67.

67

Park, D.-H., Lee, J ., & Han, I. (2007). The effect of on-line consumer reviews on
consumer purchasing intention: The moderating role of involvement.
International Journal of Electronic Commerce, 11, 125-148.

Park, D.-H., & Lee, J. (2008). eWOM overload and its effect on consumer behavioral
intention depending on consumer involvement, Electronic Commerce Research
and Applications, 7, 386-398.

Petty, R. E., Cacioppo, J. T., & Goldman, R. (1981). Personal involvement as a
determinant of argument based persuasion. Journal of Personality and Social
Psychology, 41 , 847-855.

Piller, C. (1999) Everyone is a critic in cyberspace. Los Angeles Times, December 3, A1.
Phelps, J. E., Lewis, R., Mobilio, L., Perry, D., & Rarnan, N. (2004). Viral marketing or

electronic word of mouth advertising: Examining consumer responses and
motivations to pass long email. Journal of Advertising research, 45, 333-348.

 

Pitta, D. A., & Fowler, D. (2005) Internet Community forums: an untapped resource of
consumer marketers. Journal of Consumer Marketing, 22, 265-274.

Pollach, I. (2006). Electronic word of mouth: A genre analysis of product reviews on
consumer opinion web Sites. Proceedings of the 36th Hawaii International
Conference on System Science.

Reynolds R. A. & Reynolds J. L. (2002). Evidence. In J. P. Dillard & M. Pfau (Eds) The
persuasion handbook (pp. 427-444). Thousand Oaks CA: Sage.

Reichheldm, F. F. (1996). The loyalty eflect. Cambridge, MA: Harvard Business School
Press.

Reinard, J. C. (1988). The empirical study of the persuasive effects of evidence: The
status afier fifty years of research. Human Communication Research, 15, 3-59.

Resnick, P., & Zeckhauser, R. (2002). Trust among strangers in intemet transactions:
Empirical analysis of eBay’s reputation system. In M. R. Baye (ed.), The
economics of the Internet and E- COMMERCE, Amsterdam, Elsevier Science, 11,
127-157.

Resnick, P., Zeckhauser, R., Swanson, J ., & Lockwood, K. (2006). The value of
reputation on eBay: A controlled experiment, Experimental Economics, 9, 79-101.

Reyna, V. F., Woodruff, W. J ., & Brainerd, C. J. (1987). Attitude change in adults and
adolescents: Moderation versus polarization, statistics versus case histories.
Unpublished manuscript, University of Texas, Dallas.

68

Richins, M. L. (1983). Negative word of mouth by dissatisfied customers: a pilot study.
Journal of Marketing, 4 7, 68-78.

Richins, M. L., & Root-Shaffer, T. (1988) The role of involvement and opinion
leadership in consumer word of mouth; an implicit model made explicit.
Advertising Consumer Research, 15, 32-36.

Rogers, E. M. (1995). Diflusions of Innovations (4‘h ed.). New York: The Free Press.

Ryan, B., & Gross, N. (1945). The diffusion of hybrid seed corn in two Iowa
communities. Rural Sociology, 8, 15-24.

Sen, 8. (2008) Determinants of consumer trust of virtual word-of-mouth: An observation
study from a retail Website, Journal of American Academy of Business,
Cambridge, 14, 1.

Silverman, G. (1997). How to harness the awesome power of word of mouth. Direct
Marketing, 60(7), 32-37.

Schwartz E. I. (1998). OK. Retailers, Why do your own marketing when you make
100,000 other web sites do it for you? New York Times, Aug 10, 3.

Seiter. J. 8., & Seiter, D. R. (2005). Consumer Persuasion: The Use of Evidence When
Negotiating the Price of a New Automobile, Journal of Applied Social
Psychology, 35, 4, 686-696.

Sen, 8. (2008). Determinants of Consumer Trust of Virtual Word-of-Mouth: An
Observation Study from a Retail Website, Journal of American Academy of
Business, Cambridge, 14, 1, 30-35.

Sen, 8., & Lerman, D. (2007). Why are you telling me this? An examination into
negative consumer reviews on the Web. Journal of Interactive Marketing, 21, 76-
94.

Sheth, J. N. (1971). Word of Mouth in Low-Risk Innovations. Journal of Advertising
Research, 113, 15-18.

Stiff, J. B. (1986). Cognitive processing of persuasive message cue: A meta-analytic
review of the effects of supporting information on attitudes, Communication
Monographs, 53, 75-89.

Slater, M. D., & Rouner, D. (1996). Value-affirmative and value-protective processing of
alcohol education messages that include statistical evidence or anecdotes.
Communication Research, 23 , 210-235.

69

 

Sun, T., Youn, 8., Wu, G., & Kuntaraporn, M. (2006). Online word of mouth : An
exploration of its antecedents and consequences, Journal of Computer-mediated
Communication, 11, 1 104-1 127.

Swan, J. E., & Oliver R. L. (1989). Postpurchase communications by consumers. Journal
of Retailing; 65, 516-533.

Tan, C. T., & Dolich, I. J. (1983) .A comparative study of consumer information seeking:
Singapore Versus U.S. Journal of the Academy of Marketing Science, 11, 313-322.

Taylor, 8. E., & Thompson, S. C. (1982). Stalking the elusive “vividness” effect.
Psychological Review, 89, 155-181. F

Udell, J. G. (1966). Prepurchase behavior of buyers of small electrical Appliances.
Journal of Marketing, 30, 50-52.

 

Vanderford, M. L., Smith, D. H., & Harris W. 8. (1992) Value identification in narrative '
discourse: Evaluation of a HIV education demonstration project. Journal of
Applied Communication Research, 20, 123-160.

Wagenheim, F., & Bayon, T. (2004). Satisfaction, loyalty and word of mouth within the
customer base of a utility provider: differences between stayers, switchers and
referral switchers, Journal of Consumer Behaviour, 3, 3, 211-220.

Walker, C. (1995). Word-of-mouth. American Demographics, 1 7, 38-44.

Wells, G. L, & Harvey, J. H. (1977). Do people use consensus information in making
causal attributions? Journal of personality and social Psychology, 35 , 279-293.

Wernerfelt, B. (1994). On the function of sales assistance, Marketing Science, 13, 68-82.

Williams, F., Strover, 8., & Grant, A. E. (1994). Social aspects of new media
technologies. In J. Bryant & D. Zillmann (Eds), Media eflects: Advances in
theory and research (pp. 463-482). Hillsdale: NJ: Lawrence Erlbaum Associates.

Wilson, E. J. & Sherrell, D. L. (1993) Source effects in communication and persuasion
research: A meta-analysis of effect size, Journal of the Academy of Marketing
Science, 21, 2,101-112.

Whyte, W. H., Jr. (1954). The Web of Word of Mouth, Fortune, 50, 140-143.

Ying, H. L., & Chung, C. M. Y. (2006). The effects of single-message single-source

mixed word-of-mouth on product attitude and purchase intention. Asia Pacific
Journal of Marketing and Logistics, 19, 75-86.

70

Zillmann, D., Perkins, J. W., & Sundar, S. S. ( 1992 ). Impression— formation effects of
news varying in descriptive precision and exemplifications. Medienpsychologie, 4,
168-185.

71