THE INFLUENCE OF CATEGORICAL CUES OF USER-GENERATED CONTENT ON
ATTITUDE: A TEST OF SELF-CATEGORIZATION THEORY
By
Jeong-woo Jang

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Communication - Doctor of Philosophy
2014

i

i

ABSTRACT
THE INFLUENCE OF CATEGORICAL CUES OF USER-GENERATED CONTENT ON
ATTITUDE: A TEST OF SELF-CATEGORIZATION THEORY
By
Jeong-woo Jang
User-generated content (UGC) influences viewers’ attitudes. Previous findings suggest that
various types of social categorical information of a user (e.g., race, gender) embedded in
UGC trigger categorized perceptions (in-group vs. out-group) among viewers. Based on selfcategorization theory (Turner, 1987) the current study predicts that, once categorized
perceptions take place on participatory websites, people identify with their in-group, and thus
UGC from the in-group exerts greater influence than that of an out-group. In addition, the
study proposes that prototypical expertise of categorized groups and a viewer’s involvement
with a given topic may moderate the influence of in-group UGC on viewers’ attitudes; the
gap between in-group influence and out-group influence is more pronounced when 1) the ingroup (vs. the out-group) has prototypical expertise on a topic and/or when 2) a viewer’s
involvement with a topic is high (vs. low). In testing the proposed predictions, the current
study categorized individuals by gender (male vs. female). To test the hypotheses, 276
participants viewed a mock webpage of Yelp.com, which displayed information about a local
business along with four user-generated comments. The study employed a 3 (gender-typed
topic: masculine vs. feminine vs. gender-neutral) x 3 (composition of UGC: two positive
comments from male commenters and two negative comments from female commenters vs.
two negative comments from male commenters and two positive comments from female
commenters vs. one positive comment from male commenter, one negative comment from
male commenter, one positive comment from female commenter, and one negative comment
from female commenter) between-subject design. The results were consistent with selfcategorization theory. In-group commenters exerted greater influence on viewers’ attitudes
ii

than did out-group commenters. Moreover, prototypical expertise of categorized groups and
topic involvement moderated the influence of in-group UGC, but only when there was high
in-group identification. The gap between in-group influence and out-group influence was
significant when the in-group had prototypical expertise on a topic and viewers scored high
on in-group identification. Also, the gap between in-group influence and out-group influence
was statistically meaningful when involvement was high and viewers possessed high ingroup identification.

iii

ACKNOWLEDGEMENTS

First and foremost, I thank God for leading my way.

I express my deep gratitude to my advisor, Dr. Joseph B. Walther, for his invaluable guidance,
support, and intelligence throughout my entire graduate training.

I gratefully acknowledge my doctoral committee members, Dr. Gary Bente, Dr. Maria
Lapinski, and Dr. Hairong Li, for their insightful comments and careful critique of this
project.

I thank Dr. Eun-Ju Lee for her invaluable advice and support.

I am fortunate to have the opportunity to work with amazing colleagues and friends. Their
support and encouragement led this project toward its completion.

I would like to thank my family for their unwavering support and prayer.

Finally, I thank Yoon Koo Lee for his never-ending support, encouragement, and
unconditional love throughout my project.

iv

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................................. VII
vii
INTRODUCTION ..................................................................................................................... 8
LITERATURE REVIEW ........................................................................................................ 11
Influence of UGC within Participatory Websites .................................................................... 11
Identification with In-Group and In-Group Influence ............................................................. 12
Self-Categorization Theory. ............................................................................................. 13
Moderators of In-Group Influence ........................................................................................... 15
Prototypical Expertise of Categorized Groups. ................................................................ 16
Topic Involvement............................................................................................................ 17
HYPOTHESES ........................................................................................................................ 19
METHOD ................................................................................................................................ 21
Pilot Test .................................................................................................................................. 21
Participants. ...................................................................................................................... 21
Masculinity and Femininity of a Topic. ........................................................................... 21
User-Generated Content. .................................................................................................. 24
The Main Experiment .............................................................................................................. 25
Procedure. ......................................................................................................................... 25
Participants. ...................................................................................................................... 26
Experimental Design. ....................................................................................................... 26
Stimulus Materials. ........................................................................................................... 27
Yelp webpages........................................................................................................... 27
User-generated content. ............................................................................................. 27
Manipulation Check. ........................................................................................................ 27
Perceived masculinity/femininity of a topic. ............................................................. 27
Measured Variables. ......................................................................................................... 28
Attitude toward the topic. .......................................................................................... 29
Identification with in-group commenters. ................................................................. 29
Involvement. .............................................................................................................. 30
Perceived credibility of commenters. ........................................................................ 30
Computed Variables. ........................................................................................................ 31
In-group and out-group commenters’ influence on a viewer’s attitude. ................... 31
RESULTS ................................................................................................................................ 32
DISCUSSION .......................................................................................................................... 35
Issues for Future Research ....................................................................................................... 36
APPENDICES ......................................................................................................................... 40
Appendix A .............................................................................................................................. 41
Appendix B .............................................................................................................................. 43
Appendix C .............................................................................................................................. 47
v

REFERENCES ........................................................................................................................ 50

vi

LIST OF TABLES

Table 1 Sample size for experimental conditions .................................................................... 47
Table 2 Influence of the in-group commenters on viewers’ attitudes ..................................... 48
Table 3 Influence of the out-group commenters on viewers’ attitudes ................................... 48
Table 4 Comment valence and credibility evaluation of commenters ..................................... 49

vii

INTRODUCTION
The influence of user-generated content (UGC) is now well established (see for
review, Walther & Jang, 2012). Multiple studies have found that viewers’ attitudes and
perceptions are affected by the message features of UGC, such as valence (e.g., Edwards,
Edwards, Qing, & Wahl, 2007) and argument quality (e.g., Park, Lee, & Han, 2007). Other
studies have found that viewers’ attitudes and perceptions are affected by their evaluations of
the UGC source (e.g., Willemsen, Neijens, & Bronner, 2012; Winter & Krämer, 2012). In
recording the influence of UGC on viewers’ attitudes and perceptions, prior studies mainly
focused on message-related and source-related aspects of UGC. What is in need of further
scholarly attention, however, is how viewers’ attitudes and perceptions are affected by
categorical cues embedded in UGC. For instance, a viewer is often able to gather, by looking
at the profile picture or user name, categorical information (e.g., gender, race) about the user
who generated the UGC.
Especially, categorical cues may play an important role in understanding the
influence of UGC on viewers’ attitudes and perceptions when multiple users’ comments
contradict one another. When encountering inconsistent information, people often want to
sort out the inconsistencies and determine what information is suitable for them (Weaver,
1980). The current study argues that when facing inconsistent users’ comments, viewers are
liable to turn to categorical information to sort out the inconsistent comments, which makes
certain comments more influential than others.
Categorical information, according to self-categorization theory (Turner, 1987) often
makes salient the social categories to which people belong and cause people to distinguish
their in-group from an out-group. Once people distinguish their in-group from an out-group,
the theory predicts that they identify with the in-group by assimilating themselves into the ingroup and internalizing its values. By doing so, the in-group exerts greater influence on
8

individuals’ attitudes and perceptions than does the out-group (i.e., in-group favoritism; Hogg
& Terry, 2000). Thus, if categorical cues embedded in UGC trigger categorized perceptions
and identification with an in-group, based on self-categorization theory, the current study
suggests that in-group UGC ought to exert greater influence than out-group UGC.
On participatory websites, however, the influence of in-group UGC can be
contingent upon certain conditions. First, self-categorization theory holds that categorized
perceptions (in-group vs. out-group) are often accompanied by prototypes (Turner, 1987).
Prototypes are features (e.g., attitudes, feelings, behaviors) that best characterize one category
of people and distinguish them from others (Hogg, 2011). By defining groups as distinct
entities, individuals are better able to distinguish one group from another (Hogg & Terry,
2000; Reid, Giles, & Harwood, 2005).
What might lie outside the purview of self-categorization theory is what happens
when a prototype of an out-group is perceived to be an expert. The literature on prototypical
expertise suggests that an expectation of prototypical expertise affects one’s attitude and
perception (e.g., Hollingshead & Fraidin, 2003; Huddy & Terkildsen, 1993; Lee, 2007). Thus,
the question remains open as to whether the influence of in-group UGC exceeds that of outgroup UGC even when the out-group has prototypical expertise on a topic.
In addition, previous literature indicates that involvement may play an important role
in understanding the way people process UGC (Park & Lee, 2008; Park, Lee, & Han 2007).
On participatory websites, individuals often experience information overload. They may
encounter a large amount of UGC (Park & Lee, 2008) or may find inconsistent viewpoints
within multiple users’ comments (Luo, Lan, Wang, & Ma, 2013). In dealing with information
overload, people may seek available strategies or heuristics to sort out the information (see
for review, Malhotra, 1984). In particular, previous studies found that, depending on their
level of involvement, people engage in different strategies to sort out information on
9

participatory websites. For instance, individuals with high involvement focus on the quality
of UGC, whereas individuals with low involvement may focus on the quantity of UGC (Lee,
Park, & Han, 2008; Park & Lee, 2008).
In particular, motivational theory of social identity (Hogg, 2000) posits that
depending on the level of involvement, people may make different use of the categorical cues
they find with the UGC; when an issue is important to individuals, they become motivated to
reduce uncertainty and tend to be influenced by group membership and group behavior
(Hogg, 2000; Mullin & Hogg, 1999). Based on this rationale, the current study predicts that,
depending on the level of involvement, the influence of categorical membership and in-group
UGC on viewers’ attitudes and perceptions varies. That is, when their involvement is high,
individuals rely on categorical membership and become susceptible to the influence of ingroup UGC. Taken together, the current study aims to understand how UGC, in conjunction
with various categorical cues that are embedded in UGC, affects viewers’ attitudes and
perceptions.
The study begins by discussing the influence of users’ comments on viewers’
attitudes and perceptions on participatory websites. Then the study establishes a theoretical
rationale based on self-categorization theory and in-group favoritism. Specifically, the
present study revisits self-categorization theory by testing whether 1) the categorical cues
embedded in UGC trigger categorized perceptions (in-group vs. out-group) and in turn lead to
in-group identification, and whether 2) identification with the in-group predicts in-group
influence on participatory websites. Furthermore, this study examines two potential
moderators of in-group favoritism—prototypical expertise of categorized groups as well as a
viewer’s involvement with a given topic—in order to extend our understanding of how
categorical cues embedded in UGC affect viewers’ attitudes and perceptions.

10

LITERATURE REVIEW
Influence of UGC within Participatory Websites
Participatory websites (or web 2.0 systems) present messages from multiple sources.
Specifically, participatory websites juxtapose messages that are generated by the proprietor of
a webpage, a computational system, and/or nonproprietary users (i.e., user-generated content;
see for review, Walther & Jang, 2012). For instance, messages that are generated by the
proprietor of a webpage (e.g., a video on YouTube, information about a local business on
Yelp, product information on Amazon) are followed by UGC (nonproprietary users’
comments) and by information generated by a computational system (e.g., the number of
one’s Facebook friends, the number of times users have viewed a particular content, other
site-related user behaviors). Participatory websites, such as social networking sites (e.g.,
Facebook), video-sharing sites (e.g., YouTube), local business review sites (e.g., Yelp), and
review-based commercial websites (e.g., Amazon, TripAdvisor) encourage viewers to
contribute messages to others’ webpages.
Messages on participatory websites possess different attributes and render various
types of social influence (Walther & Jang, 2012). Of particular interest are studies that have
investigated that how various aspects of UGC affect viewers’ attitudes and perceptions. In
particular, message-related characteristics of UGC, such as the valence of UGC, may
influence viewers’ attitudes and perceptions. Walther, DeAndrea, Kim, and Anthony (2010)
found that positive vs. negative comments about a public service announcement (PSA) video
affected viewers’ attitudes toward the video. When participants were exposed to negative (vs.
positive) comments concerning the PSA video about the risk of marijuana use, they evaluated
the video in a more negative (vs. positive) light. In reference to the valence of UGC, Edwards
et al. (2007) similarly found that viewing negative (vs. positive) comments about a college

11

professor on a professor rating website resulted in less professor credibility and less
attractiveness.
In addition to valence, previous studies have identified various message-related
aspects of UGC that affect viewers’ attitudes and perceptions. For instance, viewers’ attitudes
and perceptions about a news article were found to be affected by congruency between the
news article and UGC (Lee, Jang, & Kim, 2009). Viewing others’ comments that are
incongruent with the tone of the news article led to less positive evaluations of the article. On
product review sites, as the proportion of negative product reviews increased, participants
reported more negative attitudes toward the product (Lee, Park, & Han, 2008).
In addition to message-related characteristics, previous literature reveals that viewers’
attitudes and perceptions can be affected by how they evaluate the individuals who generated
the comments. Willemsen et al. (2012) found that perceived expertise and trustworthiness of
a UGC source predicted viewers’ attitudes toward the UGC. When a source proclaimed
his/her expertise on a topic, viewers’ attitudes toward the UGC (e.g., bad vs. good, not useful
vs. useful) were affected by viewers’ perceptions of the source’s expertise; if viewers
perceived the source to be trustworthy, their attitudes toward the UGC were changed by their
perceptions of the source’s trustworthiness. Winter and Krämer (2012) also found that UGC
created by an expert source was likely to be selected more frequently, to be read longer, and
to be evaluated in a more positive way, than UGC created by a non-expert source.
Prior studies have identified various features of UGC that affect viewers’ attitudes
and perceptions. In attempting to further extend our understanding of how UGC leads to
social influence, the current study focuses on the effects of categorical cues embedded in
UGC on viewers’ attitudes and perceptions.
Identification with In-Group and In-Group Influence

12

Participatory websites often contain various categorical cues about users (e.g., gender,
race; Ren et al., 2012; Walther & Jang, 2012). For instance, a profile picture of an individual
wearing a shirt with a university’s emblem on it makes salient the social category of school
affiliation (Ganster, 2011). User names alone can prompt categorized perceptions pertaining
to race or gender (Jang & Walther, 2013).
Understanding the role of categorical cues embedded in UGC is crucial especially
when multiple users’ comments about an object or topic contradict one another. When
encountering inconsistent comments, the current study suggests that a viewer may organize
them according to his categorization of himself and the various commenters. When certain
comments are categorized as having come from the viewer’s in-group, the in-group
comments may exert persuasive influence.
Self-Categorization Theory.
Self-categorization theory (Turner, 1987) assumes that the way people perceive
themselves can range from being unique individuals (personal identity) to belonging to social
groups or wider social categories, such as gender, ethnicity, and occupations (social identity).
According to Turner (1986, 1987), when personal identity is salient, people perceive
themselves as idiosyncratic individuals. Conversely, when social identity becomes strong,
people differentiate one group from another, based on shared and unshared social categories.
In turn, as social identity strengthens, people tend to identify themselves with ingroup members (Turner, 1987; Turner & Oakes, 1989). People assimilate themselves with ingroup members by defining themselves as representatives or exemplars of the social category
to which they belong. When people identify themselves with their in-group members, they
become susceptible to the influence of other members of their social groups or social
categories since they internalize the perspectives or norms of the group to which they belong
(Hornsey, 2008; Turner, 1987; Voci, 2006).
13

Self-categorization theory posits that social identity is a function of accessibility and
fit (Oakes, Turner, & Haslam, 1991; Turner, 1987). Accessibility refers to the readiness with
which a stimulus input with given properties becomes identified as a category (Bruner, 1957,
p. 133; Turner, 1987). The theory argues that accessibility is necessary to invoke or activate
categorization among individuals. Fit, on the other hand, refers to the extent to which a given
categorization matches social reality (Turner, 1987). That is, categorical distinction needs to
be consistent with social reality, so the given categorization can appear to be a reasonable
way of organizing the world (Turner, 1991).
Self-categorization theory holds that fit operates according to the metacontrast
principle. The metacontrast principle maximizes perceived inter-category differences and
minimizes intra-category differences (Turner, 1987). When perceived differences between ingroup and out-group are maximized and perceived differences between in-group members are
minimized, the theory argues that people tend to perceive their social category as a single unit,
entity, or group. Thus, people tend to perceive the given categories as a reasonable way to
organize the world (Oakes, Turner, & Haslem, 1991; Simon, Hastedt, & Aufderheide, 1997).
Moreover, the metacontrast principle leads to categorized perceptions being
accessible (Reid & Hogg, 2005; Turner, 1987; Turner & Oakes, 1989). For instance, when
group category matches the position being advocated by a group (e.g., Group A supports
Position 1 and Group B supports Position 2), people attribute the difference between the
groups to group category. That is, when inter-category differences are maximized and intracategory differences are minimized as the metacontrast principle suggests, the differences
between the groups are attributed to category-based differences and categorized perceptions
(in-group vs. out-group) became accessible.
Through the metacontrast principle, the theoretical postulates of self-categorization
theory can be applied to the context of participatory websites. Jang and Walther (2013) found
14

that categorical cues embedded in UGC led to categorized perceptions among viewers and
caused them to identify with their in-group. In their study, participants viewed four users’
comments: two comments were displayed with Caucasian profile pictures with typical
Caucasian names and the other two were shown with Asian profile pictures with typical
Asian names. Based on the metacontrast principle, between-group differences were
maximized and within-group differences were minimized to induce salient social identity.
Consistent with what self-categorization theory predicts, the study found that Asian
participants identified themselves with the Asian commenters and Caucasian participants
identified with the Caucasian commenters.
The current study aims to test whether self-categorization theory explains the
influence of categorical cues embedded in UGC on viewers’ attitudes. It tests whether
categorization of gender on participatory websites induces categorized perceptions among
users, makes participants identify with their in-group, and also predicts greater influence from
the in-group than the out-group. Previous studies suggest that even minimal cues can induce
distinct categorization (Dehl, 1990; Fazio & Dunton, 1997; Hewstone, Hantzi, & Johnston,
1991; Kurzban, Tooby, & Cosmides, 2001), and gender is known to be a strong predictor of
social categorization (Cornetto & Nowak, 2006; Stangor, Lynch, Duan, & Glass, 1992).
Given that various cues on participatory websites, such as profile pictures and user names,
indicate gender of users, gender-based identity may become salient on participatory websites.
Moderators of In-Group Influence
The influence of in-group UGC on viewers’ attitudes and perceptions, however, can
be contingent upon certain moderators. The current study suggests that prototypes of
categorized groups and a viewer’s involvement with a given topic may serve as moderators
that affect the influence of in-group UGC on viewers’ attitudes and perceptions. The

15

moderating roles of both prototypical expertise of categorized groups and topic involvement
are delineated below.
Prototypical Expertise of Categorized Groups.
According to self-categorization theory, when people perceive a group as a whole due
to categorization, they often develop a prototype, a feature or a position that represents the
group as a whole (Hogg & Tindale, 2005; Reid, Giles, & Harwood, 2005). Examples of such
prototypes are Asians are good at math or men are more mechanical than women. Prototypes
involve the metacontrast principle; they facilitate maximizing inter-group differences and
minimizing within-group ones, consequently making groups into distinct entities.
Interestingly, when a prototypical belief about an out-group is associated with greater
knowledge or expertise about a given topic, the influence of the in-group may not override
that of the out-group. Previous studies suggest that the expectation of prototypical expertise
affects one’s attitude and perception (e.g., Hollingshead & Fraidin, 2003; Huddy &
Terkildsen, 1993; Lee, 2007; Spencer, Steele, & Quinn, 1999; Stone, Lynch, Sjomeling, &
Darley, 1999). Lee (2007) shows that people develop prototypical expertise expectations
using categorical cues in a computer-mediated interaction context. In her study, she
arbitrarily assigned to participants sex-marked avatars and sex-typed topics and let them
interact with an ostensible partner. She found that people inferred their partners’ sex through
minimal and arbitrary cues (self-proclaimed expertise in sex-typed topics or randomly
assigned sex-marked avatars) and, based on their inference, evaluated their interactants’
opinions. For instance, when a participant inferred that his/her interactant was female, the
participant was more likely to accept the interactant’s opinion when the given topic was
associated with females. This study suggests that an expectation of expertise that stems from
prototypes or stereotypes of social categories can affect one’s attitude and perception.
Hollingshead and Fraidin (2003) also found that people used gender-based prototypes to infer
16

expertise or knowledge about a topic. Participants expected greater knowledge about
masculine topics from males and expected greater knowledge about feminine topics from
female.
Taken together, the literature suggests that viewers’ attitudes and perceptions can be
affected by having expectations of prototypical knowledge or expertise. Based on previous
findings, the current study predicts that prototypical expertise of categorized groups
moderates the influence of in-group UGC on viewers’ attitudes and perceptions. Particularly,
in-group influence is more pronounced when the in-group, and not the out-group, has
prototypical expertise on a given topic. Conversely, when a prototypical belief about an outgroup is associated with greater knowledge or expertise about a given topic, the influence of
the in-group may not override that of the out-group and accordingly, in-group influence is
likely to be less pronounced.
Topic Involvement.
Involvement plays an important role in understanding the influence of UGC on
viewers’ attitudes and perceptions. Previous studies argue that, depending on their level of
involvement, people engage in different strategies to organize and sort out information on
participatory websites. One study found that individuals with high involvement attended to
UGC content and considered the quality of its arguments, whereas individuals with low
involvement focused on the quantity of UGC when they reviewed user’ comments to make a
purchase decision (Park, Lee, & Han, 2007). Another study found that when individuals were
highly involved, the perceived informativeness of UGC predicted their purchasing intention.
However when they were less involved, perceived product popularity affected individuals’
purchasing intention (Park & Lee, 2008). These studies indicate that depending on the level
of involvement, different aspects of UGC—its quality or quantity—become influential.

17

Departing from the comparison between quality and quantity of UGC, the current
study suggests that, depending on the level of involvement, what may vary is the influence of
categorical cues embedded in UGC on individuals’ attitudes and perceptions. Specifically,
the influence of in-group UGC on viewers’ attitudes and perceptions can be more pronounced
when involvement is high. The motivational theory of social identity processes suggests that
when an issue is important or relevant to individuals, they become motivated to reduce
uncertainty and consequently rely on group membership and group behavior (Hogg, 2000).
Mullin and Hogg (1999) found that when an issue was perceived to be important, individuals
attended to categorical membership, becoming more susceptible to in-group influence. When
the issue was less important to individuals, however, they were less swayed by categorical
membership and in-group.
Based on this rationale, the current study predicts that the influence of in-group UGC
on viewers' attitudes and perceptions is greater when viewers have high rather than low
involvement. While the type of involvement can vary, the current study focuses on topic
involvement (or issue involvement), which refers to the extent to which a topic is of personal
importance (Petty & Cacioppo, 1979). Topic involvement motivates one to evaluate
information extensively and reduce potential uncertainty (e.g., Laurent & Kapferer, 1985;
Lee, Herr, Kardes, & Kim, 1999). Thus when topic involvement is high, individuals are likely
to rely on group membership as the motivational theory of social identity processes predicts.
Accordingly, the present study predicts that varying degrees of topic involvement may
moderate the extent to which group membership and group behavior affect individuals'
attitudes and perceptions.

18

HYPOTHESES
The current study proposes hypotheses to test whether 1) categorical cues embedded
in UGC lead to in-group identification as well as in-group favoritism and whether 2)
prototypical expertise of categorized groups and a viewer’s involvement with a given topic
moderate in-group favoritism on participatory websites. Given that gender is a well established predictor for social categorization (Cornetto & Nowak, 2006; Stangor, Lynch,
Duan, & Glass, 1992) and there are often multiple cues (e.g., profile pictures, user names)
that indicate the gender of a user posting UGC on participatory websites (Spottswood,
Walther, Holmstrom, & Ellison, 2013), the present study focuses on gender-based
categorization to test the following predictions.
First, self-categorization theory (Turner, 1987) predicts that the in-group exerts
greater influence than the out-group. Based on the theory, the current study suggests that
people form categorized perceptions through various UGC-embedded categorical cues. Once
people distinguish their in-group from the out-group, they identify with the in-group and
become susceptible to its influence. Thus, the current study predicts that regardless of the
topic in-group influence takes place.
H1: In-group (same gender) commenters affect viewers’ attitudes more than do outgroup commenters.
Second, given that self-categorization theory posits that in-group influence is a
function of in-group identification, the current study predicts that the difference between ingroup and out-group influence is more pronounced among those who score high, rather than
low, on in-group identification measures.
H2: The difference between in-group influence and out-group influence is more
pronounced with high, rather than low, in-group identification.

19

The current study also predicts that prototypical expertise of categorized groups and
topic involvement may moderate in-group favoritism on participatory websites. Previous
literature on prototypical expertise shows that when a topic triggers viewers’ expectation of
prototypical expertise, viewers are affected by those message sources who are prototypically
associated with expertise and knowledge on the topic. Based on previous findings, the present
study argues that when the in-group has prototypical expertise on a given topic, the in-group
should exert greater influence than the out-group. However, when prototypical expertise is
associated with the out-group, in-group favoritism on participatory websites becomes
questionable. Based on this notion, prototypical expertise may moderate in-group favoritism,
such that the difference between in-group and out-group influence is more pronounced when
the in-group, rather than the out-group, has prototypical expertise on the topic.
H3: The difference between in-group influence and out-group influence is more
pronounced when an in-group, rather than an out-group, has prototypical expertise on
a topic.
Lastly, depending on the level of topic involvement, the degree to which in-group
UGC affects viewers’ attitudes and perceptions may vary. Specifically, this study predicts
that the influence of categorical cues embedded in UGC on viewers’ attitudes and perceptions
is greater when viewers have high, rather than low, involvement with a given topic.
H4: The difference between in-group influence and out-group influence is more
pronounced when individuals have high, rather than low, involvement toward a topic.

20

METHOD
Participants viewed a mock webpage of Yelp.com. Yelp is one of the most popular
participatory websites for sharing information about local businesses, such as restaurants,
beauty salons, and automotive repair shops. Users simply rate businesses or write comments
to share their experiences. Motives to use Yelp vary from information seeking to
entertainment to just passing time (Hicks et al., 2012). The Yelp page mock-ups in the
experiment contained basic information about one local business as well as four comments
from users.
Pilot Test
The purpose of the pilot test was two-fold: assessing the masculinity and femininity
of several local businesses and selecting users’ comments for the main study. Participants
were instructed to assess nine local businesses in terms of their masculinity and femininity.
Then, participants assessed 10 user-generated comments that were sampled from Yelp.com in
terms of valence, helpfulness, argument quality, masculinity, and femininity. Based on the
pilot test, the main experiment selected three target businesses (masculine, feminine, genderneutral) and four users’ comments (two positive and two negative).
Participants.
A total of 50 participants were recruited for the pilot test. Approximately 42.0% of
participants were male and the sample included 72% White/Caucasian, 10% Black/AfricanAmerican, 8% Asian, 2% Hispanic/Latino, and 4% “other.” The mean age of the participants
was 20.48 (SD = 1.91), ranging from 18 to 27. Participation was voluntary and all
participants received course credit in exchange for their participation.
Masculinity and Femininity of a Topic.
A masculine topic refers to a topic that is stereotypically perceived as male-oriented
or a topic that is associated with a typical male rather than a typical female. A feminine topic
21

refers to a topic that is stereotypically perceived as female-oriented or a topic that is
associated with a typical female rather than a typical male (Carly, 2001). To come up with
one masculine topic, one feminine topic, and one gender-neutral topic, nine types of
business—an automotive repair shop, a cooking class, a coffee shop, a restaurant, an
electrical service center, a pet store, a musical instrument store, a daycare center, and a
fabrics shop—were assessed in terms of their masculinity and femininity. These types of
business are likely to be perceived as either masculine, feminine, or gender-neutral (Franck &
Rosen, 1949). Participants evaluated the masculinity and femininity of each local business on
a 9-point semantic differential scale (Hamilton & Nowak, 2010). Given that masculinity and
femininity are independent constructs, they were measured separately (Heilburn, 1976). Items
were anchored by “not masculine at all vs. very masculine,” “not male-oriented at all vs. very
male-oriented,” “not feminine at all vs. very feminine,” and “not female-oriented at all vs.
very female-oriented.” In addition to these four items, five additional items that also evaluate
other aspects of a local business were included as fillers to make less distinguishable the
purpose of this measurement. Additional items included “not professional at all vs. very
professional,” “not appealing at all vs. very appealing,” “not favorable at all vs. very
favorable,” “of poor quality vs. of good quality,” and “not pleasant at all vs. very pleasant”
(Holbrook & Batra, 1987; Mitchell & Olson, 1981). Fillers were not further analyzed (see
Appendix A).
For the main experiment, the study selected as the masculine topic a business
perceived to be highly masculine while also featuring a negative correlation between its
perceived masculinity and its perceived femininity. As the feminine topic, the study selected
a business perceived to be highly feminine while also featuring a negative correlation
between its perceived masculinity and its perceived femininity. Lastly, as the gender-neutral

22

topic, the study selected a business with equivalent levels of perceived masculinity and
perceived femininity.
Perceived masculinity was measured by averaging the two items, “not masculine at all
vs. very masculine” and “not male-oriented at all vs. very male-oriented.” Perceived
femininity was measured by averaging the two items, “not feminine at all vs. very feminine”
and “not female-oriented at all vs. very female-oriented.” The masculine business (an electric
service company) for perceived masculinity ranged from 2 to 9 on a 9-point scale, α = .87, M
= 7.34, SD = 1.46 and for perceived femininity ranged from 1 to 6, α = .88, M = 2.89, SD =
1.26. The bivariate correlation between masculinity and femininity of the masculine business
was r (48) = -.58, p < .001. The feminine business (a fabric shop) for perceived masculinity
ranged from 1 to 6 on a 9-point scale, α = .83, M = 2.32, SD = 1.24 and for perceived
femininity ranged from 4 to 9, α = .81, M = 7.48, SD = 1.24. The bivariate correlation
between masculinity and femininity of the feminine business was r (48) = -.70, p < .001. The
gender-neutral business (a musical instrument store) for perceived masculinity ranged from 3
to 9 on a 9-point scale, α = .85, M = 5.57, SD = 1.11 and for perceived femininity ranged
from 3 to 9, α = .81, M = 5.28, SD = 0.95. The bivariate correlation between masculinity and
femininity of the gender-neutral business was r (48) = .19, p = .19.
A single-factor within-subject ANOVA was conducted to see if the masculinity and
femininity of the chosen businesses differed significantly from one another. The results
revealed a significant difference in perceived masculinity across three topics, F (2, 147) =
164.35, p < .001. Post hoc comparisons using Scheffe’s procedure (p < .05) showed that the
masculine topic was perceived to be more masculine than the feminine and gender-neutral
topics. The results also found a significant difference in perceived femininity across three
topics, F (2, 147) = 187.76, p < .001. Post hoc tests using Scheffe’s procedure (p < .05)
showed that the feminine topic was perceived to be more feminine than the masculine and
23

gender-neutral topics. Thus, for the main study, the most masculine (an electric service
company), the most feminine (a fabric shop), and one gender-neutral target object (a musical
instrument store) were employed.
User-Generated Content.
A total of 10 user-generated comments were sampled from Yelp.com. Participants
were asked to review the comments carefully and rate the comments in terms of valence,
helpfulness, argument quality, masculinity, and femininity. Valence was measured in order to
select the two most positive and the two most negative comments for the main experiment.
Helpfulness and argument quality were measured to avoid any potential confounding
influence. If comments vary in these aspects, the influence of UGC on viewers’ attitudes
could be a function of helpfulness or of argument quality (Chu & Kamal, 2008; Racherla &
Friske, 2012). Lastly, each review was measured in terms of how masculine and feminine it
appeared. Given that each review could be associated with both male and female commenters,
the main experiment used a review that appeared neither overtly masculine nor feminine.
Accordingly, the main study employed the two most positive and the two most negative usergenerated comments, which had equivalent levels of helpfulness and of argument quality,
without any gender-typicality.
Participants evaluated each of the comments on a 7-point Likert scale (1 = very
negative, 7 = very positive for valence; 1 = not helpful at all, 7 = very helpful for helpfulness;
1 = very weak, 7 = very strong for argument quality, 1 = not masculine at all, 7 = very
masculine for masculinity, 1 = not feminine at all, 7 = very feminine for femininity). The two
most positive reviews ranged from 2 to 7, M = 6.40, SD = 0.94 for valence, ranged from 2 to
7, M = 5.66, SD = 1.05 for helpfulness, ranged from 1 to 7, M = 5.41, SD = 1.15 for argument
quality, ranged from 2 to 6, M = 3.92, SD = 0.70 for masculinity, and ranged 2 to 6, M = 3.99,
SD = 0.81 for femininity. The two most negative reviews ranged from 1 to 4, M = 1.44, SD =
24

0.75 for valence, ranged from 1 to 7, M = 5.11, SD = 1.61 for helpfulness, ranged from 1 to 7,
M = 4.88, SD = 1.60 for argument quality, ranged from 2 to 7, M = 3.83, SD = 0.85 for
masculinity, and ranged from 2 to 7, M = 3.80, SD = 0.66 for femininity.
A single-factor within-subject ANOVA was conducted to see if the valence of UGC
differed significantly from one another. The results showed that the four users’ comments
significantly differed in their valence, F (3, 144) = 403.06, p < .001. Post hoc comparison
using Scheffe’s procedures (p < .05) showed that the two positive comments (M = 6.30a, SD
= 1.13 and M = 6.50a, SD = 0.89) were perceived to be more positive than were the negative
comments (M = 1.53b, SD = 0.84 and M = 1.33b, SD = 0.75). There was no difference in
perceived valence between the two positive comments or between the two negative
comments.
Used for the main study were the two most positive and two most negative usergenerated comments with equivalent levels of helpfulness, argument quality, masculinity, and
femininity. Positive comments consisted of, “They are the best around the town. I strongly
recommend this place” and “I have been with this place since college and will never leave it.”
Negative comments included, “So far, the experience has been terrible. I would never
recommend this place” and “This place even does not deserve stars. If I could give them
negative stars, I would.”
The Main Experiment
Procedure.
In an online experiment, participants viewed one of nine mock webpages of Yelp.com.
Each experimental webpage included basic information about a local business, such as its
location, webpage address, and pictures of the business, as well as four user-generated
comments about it. Participants were instructed to take enough time to view the information
about a local business as well as the users’ comments posted on Yelp.com. Participants then
25

answered dependent measure questionnaires (see Appendix B). Participation was voluntary
and all participants received course credit in exchange for their participation.
Participants.
A total of 302 undergraduate students participated in this study via an online
participant pool at two large universities located in the Midwest: 242 students from one
university and 60 students from another. To facilitate rapid participation, the current study
employed the convenient sampling approach (Yu & Cooper, 1983). The students from the
two universities did not differ in their demographics and responses for the dependent
measures, so the data from the two universities were collapsed. Data from 26 participants
were dropped because the participants failed to complete the questionnaire or responded
carelessly (e.g., indicating the same value for all questionnaire items). Thus, for the rest of the
analyses, the study used data from 276 participants. Participants were randomly assigned to
one of the nine experimental conditions, so approximately 31 participants were assigned to
each condition (see Table 1).
Approximately 43.5% of participants were male, and the sample included 76.1%
White/Caucasian, 8.0% Black/African-American, 11% Asian, 1.1% Hispanic/Latino, and 3.6%
“other.” The mean age of the participants was 21.84 (SD = 4.17), ranging from 18 to 50.
Experimental Design.
A 3 x 3 between-subject design was used. The first variable—
gender-typed topic—had three levels: masculine (an electric service company), feminine (a
fabrics store), and gender-neutral (a musical instrument store). The second variable—
composition of user-generated comments—also had three levels. Participants viewed either 1)
two positive comments from male users and two negative comments from female users, 2)
two negative comments from male users and two positive comments from female users, or 3)
one positive comment from a male user, one positive comment from a female user, one
26

negative comment from a male user, and one negative comment from a female user). The
experimental design allowed the current study to test whether 1) predictions drawn from selfcategorization theory and in-group favoritism occurred and whether 2) prototypical expertise
of categorized groups and topic involvement moderated in-group favoritism on participatory
websites.
Stimulus Materials.
Yelp webpages. Mock webpages of Yelp contained information about a local business
(an electric service company, a fabric shop, or a musical instrument store) and four comments
from users.
User-generated content. Participants viewed a total of four user-generated comments
juxtaposed with information about the target business. Out of the four user-generated
comments, two comments appeared to come from male commenters and two from female
commenters. In order to manipulate self-categorization, a comment that was supposed to have
been generated by a male commenter was shown with a picture of a male along with a
common male name (Mike, Steve). A comment that was supposed to be generated by a
female commenter was shown with a picture of a female along with a common female name
(Julia, Amanda). The profile pictures of the two male commenters and the two female
commenters were sampled from Yelp.com.
Drawing on the metacontrast principle, social category (i.e., gender of the
commenters) was matched with the position of the comments to create distinct in-group vs.
out-group perceptions. Specifically, the two commenters of the same gender expressed
similar opinions. Conversely, male and female commenters held opposing opinions.
Manipulation Check.
Perceived masculinity/femininity of a topic. Perceived masculinity and femininity of a
target business was measured by using the same four items used in the pilot test. Participants
27

rated perceived masculinity and femininity of each business on a 7-point Likert scale (1 =
strongly disagree, 7 = strongly agree). The reliabilities and means of this measure were, for
perceived masculinity, α = .85, M = 4.98, SD = 1.89 and, for perceived femininity, α = .93, M
= 5.15, SD = 1.86.
The masculine business ranged, for perceived masculinity, from 2 to 9 on a 9-point
scale, α = .87, M = 6.48, SD = 1.55 and, for perceived femininity, it ranged from 1 to 7, M =
3.69, SD = 1.56. The bivariate correlation between masculinity and femininity of the
masculine business was r (88) = -.55, p < .001. The feminine business ranged, for perceived
masculinity, from 1 to 7 on a 9-point scale, M = 2.82, SD = 1.44 and, for perceived femininity,
from 2 to 9, M = 6.82, SD = 1.53. The bivariate correlation between masculinity and
femininity of the feminine business was r (87) = -.51, p < .001. The gender-neutral business
ranged, for perceived masculinity, from 1 to 9 on a 9-point scale, M = 5.38, SD = 1.26 and,
for perceived femininity, it ranged from 2 to 9, M = 4.70, SD = 1.27. The bivariate correlation
between masculinity and femininity of the gender-neutral business was r (95) = - .03, p = .80.
A single-factor between-subject ANOVA was conducted to see if the masculinity and
femininity of each topic differed substantially from each other. The results revealed a
significant difference in perceived masculinity across three topics, F (2, 273) = 157.15, p
< .001. Post hoc comparisons using Scheffe’s procedure (p < .05) showed that the masculine
topic was perceived to be more masculine than the feminine and gender-neutral topics. The
results also found a significant difference in perceived femininity across three topics, F (2,
273) = 108.27, p < .001. Post hoc tests using Scheffe’s procedure (p < .05) indicated that the
feminine topic was perceived to be more feminine than the masculine and gender-neutral
topics.
Measured Variables.

28

Attitude toward the topic. Attitude toward the target business was assessed using six
items measured on a 9-point semantic differential scale anchored by “dislike vs. like,”
“positive vs. negative,” “favorable vs. unfavorable,” “bad vs. good,” “of poor quality vs. of
good quality,” and “appealing vs unappealing” (Holbrook & Batra, 1987; Mitchell & Olson,
1981). Two additional items also measured attitudes toward the target business on a 9-point
Likert scale. The items were: “How would you rate your overall evaluation of this business?”
and “How would you rate the quality of service this business offers?” Overall, the measures
were reliable, α = .93, M = 4.73, SD = 1.23.
Identification with in-group commenters. Identification with in-group commenters
was measured on a 7-point Likert scale using 11 items ranging from 1 (strongly disagree) to 7
(strongly agree), α = .82, M = 4.96, SD = 0.78 (adapted from Leach et al., 2008; Schmader,
2002). Participants were instructed to indicate the extent to which they agreed or disagreed
with the items that followed. Male participants answered “Being a man is an important part of
my self-image,” “Being a man is important to my sense of what kind of person I am,” “Being
a man is an important reflection of who I am,” “Being a man has very little to do with how I
feel about myself,” “I have a lot in common with the average man,” “I am similar to the
average man,” “I feel a bond with men,” “Men have a lot in common with each other,” “I
identify with other men,” “Being a man feels natural to me,” and “I feel personally implicated
when men are criticized.” Female participants answered “Being a woman is an important part
of my self-image,” “Being a woman is important to my sense of what kind of person I am,”
“Being a woman is an important reflection of who I am,” “Being a woman has very little to
do with how I feel about myself,” “I have a lot in common with the average woman,” “I am
similar to the average woman,” “I feel a bond with women,” “Women have a lot in common
with each other,” “I identify with other women,” “Being a woman feels natural to me,” and “I
feel personally implicated when women are criticized.”
29

Involvement. The study measured involvement with the target business (topic).
Individuals may vary in their involvement with a topic, which may affect how they process
categorical cues embedded in UGC. Involvement was measured on a 7-point Likert scale
using five items (adapted from Cho & Boster, 2005) ranging from 1 (strongly disagree) to 7
(strongly agree), α = .81, M = 4.05, SD = 1.32. The items included “The topic has little
impact on my life,” “My quality of life would not change depending on the topic,” “The topic
affects my life,” “All in all, the effects of the topic on my life would be little,” and “It is easy
for me to think how the topic influences my well-being.”
Perceived credibility of commenters. Dual process models, such as the elaboration
likelihood model (Petty & Cacioppo, 1979), argue that as their levels of involvement vary,
individuals can be influenced by either argument quality or source characteristics (Petty &
Cacioppo, 1979). Since the argument quality of each comment was already controlled
through the pilot test, such that all four users’ comments used in the main experiments had
equivalent levels of argument quality, F (3, 196) = 1.08, p = .36, and perceived helpfulness, F
(3, 196) = 0.86, p = .46, participants only answered how they would evaluate the credibility
of each commenter; if the commenters differ in their perceived credibility, it may affect
viewers’ evaluations on UGC. Participants viewed the comments of the four commenters
along with the commenters’ profile pictures and rated the perceived credibility of each
commenter on a 7-point semantic differential scale. The items included “unintelligent vs.
intelligent,” “untrained vs. trained,” “dishonest vs. honest,” “untrustworthy vs. trustworthy,”
“not an expert vs. expert,” “dishonorable vs. honorable,” “immoral vs. moral,” “incompetent
vs. competent,” “unethical vs. ethical,” and “not understanding vs. understanding” (adapted
from McCroskey & Teven, 1999). The reliabilities and means of credibility measure were α =
.92, M = 4.31, SD = 0.71 for female commenters and α = .93, M = 4.38, SD = 0.70 for male
commenters.
30

Computed Variables.
In-group and out-group commenters’ influence on a viewer’s attitude. The impact of
in-group UGC on a viewer’s attitude was assessed by calculating congruency between the
viewer’s position (i.e., attitude) and his/her in-group commenters’ position, M = 5.19, SD =
1.52 (see Table 2). The impact of out-group UGC on a viewer’s attitude was assessed by
calculating congruency between the viewer’s position and his/her out-group commenters’
position, M = 4.81, SD = 1.52. (see Table 3). For instance, when in-group commenters
expressed positive attitudes about the target business, then the more positive a viewer’s
attitude was, the stronger the impact of the in-group. When in-group commenters expressed
negative attitudes about a topic, then the more positive a viewer’s attitude was, the weaker
the influence of the in-group. As another example, when a female viewer read two negative
remarks about the target business from female commenters, then the more negatively the
viewer rated the place (i.e., the lower the score), the greater the in-group influence.

31

RESULTS
Based on self-categorization theory, H1 predicts that in-group commenters more
greatly affect viewers’ attitudes than do out-group commenters. In order to test H1, a pairedsample t-test was used to compare the influence of the in-group (M = 5.19, SD = 1.52) with
that of the out-group (M = 4.81, SD = 1.52). The results indicated that, across all topics, the
in-group exerted greater influence than the out-group, t (183) = 1.70, p = .045 (1-tailed). Thus,
the results were consistent with H1.
Given that self-categorization theory argues that in-group influence is a function of
in-group identification, H2 predicts that the difference between in-group influence and outgroup influence is more pronounced among those with high rather than low in-group
identification. In order to test H2, a paired-sample t-test was used to compare in-group
influence among those who scored higher on the in-group identification measure with ingroup influence among those who scored lower on the in-group identification measure. The
cutoff for the median split was 5.00 while the means for high and low in-group identification
groups were M = 5.53, SD = 0.48 and M = 4.15, SD = 0.74, respectively. The results revealed
that the in-group exerted greater influence than the out-group only when in-group
identification was high, t (103) = 2.23, p = .014 (1-tailed). When in-group identification was
low, however, in-group influence did not differ from out-group influence, t (183) = 0.17, p
= .43 (1-tailed). Thus, the results were consistent with H2.
Hypothesis 3 proposes that prototypical expertise of categorized groups moderates
the influence of in-group UGC on viewers’ attitudes. Specifically, H3 predicts that the
difference between in-group influence and out-group influence is more pronounced when the
in-group, rather than the out-group, has prototypical expertise on a topic. A paired-sample ttest was used to test H3. The results were not consistent with H3; the difference between ingroup influence, M = 5.10, SD = 1.69, and out-group influence, M = 4.97, SD = 1.69, was not
32

significant when the in-group had prototypical expertise on a topic, t (57) = 0.31, p = .38 (1tailed). When the out-group has prototypical expertise on a topic, the difference between ingroup influence, M = 5.08, SD = 1.36, and out-group influence, M = 4.92, SD = 1.36, was not
significant, t (60) = 0.47, p = .32 (1-tailed), either.
Given that the difference between in-group influence and out-group influence was
only observed among those with high in-group identification, further analysis was conducted.
Specifically, a series of paired-sample t-tests were conducted to see how prototypical
expertise of categorized groups and in-group identification affected in-group vs. out-group
influence. The results revealed that exerting greater influence was the in-group, M = 5.56, SD
= 1.71, over the out-group, M = 4.56, SD = 1.74, when the in-group had prototypical
expertise on the topic and viewers also had high in-group identification, t (33) = 1.72, p
= .047 (1-tailed). The difference between in-group influence and out-group influence was not
significant when the out-group had prototypical expertise on a topic or when the in-group had
prototypical expertise but viewers scored low on the in-group identification measure.
Lastly, H4 predicts that the difference between in-group and out-group influence is
more pronounced when individuals have high, rather than low, involvement with a topic. In
order to test H4, a paired-sample t-test was used to compare in-group influence among those
who scored higher on the involvement measure with among those who scored lower on the
involvement measure. The cutoff for the median split was 3.90 while the means for high and
low involvement groups were M = 4.87, SD = 0.84 and M = 2.78, SD = 0.82, respectively.
The results were not consistent with H4. The difference between in-group influence, M =
5.20, SD = 1.43, and out-group influence, M = 4.80, SD = 1.43, was not significant when
involvement was high, t (111) = 1.45, p = .08 (1-tailed). When involvement was low, the
difference between in-group influence, M = 5.18, SD = 1.65, and out-group influence, M =
4.82, SD = 1.65, was not significant, t (71) = 0.93, p = .18 (1-tailed).
33

Additionally, given that the difference between in-group influence and out-group
influence was only observed among those with high in-group identification, further analysis
was conducted. The results lend partial support for H4; the in-group, M = 5.40, SD = 1.60,
exerted greater influence than the out-group, M = 4.60, SD = 1.60, when viewers were highly
involved with a topic and when they also had high in-group identification, t (64) = 2.02, p
= .024 (1-tailed). The difference between in-group influence and out-group influence
disappeared when topic involvement was low or when involvement was high but viewers
scored low on in-group identification.
Post hoc Analysis
Post hoc analyses found that the valence of comments affected viewers’ evaluations
of the commenters. Positive comments led to greater perceived source credibility than did
negative comments. The results indicated that commenters who left positive comments were
perceived to be more credible than commenters who left negative ones. A series of pairedsample t-tests indicated that positive commenters were perceived to be, in general, more
credible than negative commenters (see Table 4).

34

DISCUSSION
In testing how social categorical cues embedded in UGC affect viewers’ attitudes
and perceptions, the current study yields support of self-categorization theory. Regardless of
the topic, viewers were more influenced by the in-group commenters than by the out-group
commenters. Moreover, when viewers were divided into two groups—high vs. low in-group
identification—in-group influence was observed only among those with high in-group
identification. This finding is consistent with self-categorization theory which posits that ingroup influence is a function of in-group identification. The current study also found that the
influence of in-group UGC on viewers’ attitudes and perceptions was moderated both by the
prototypical expertise of categorized groups and involvement with a given topic. However,
this occurred only when viewers had high in-group identification. Specifically, the in-group
exerted greater influence than did the out-group when in-group identification was high and
the in-group also had prototypical expertise on a topic. Likewise, when in-group
identification was high, those who found a topic to be involving were affected by in-group
UGC.
The results indicate that in-group identification plays a crucial role in understanding
the influence of categorical cues on viewers’ attitudes and perceptions. First, categorical cues
embedded in UGC caused, among those with high in-group identification, categorized
perceptions (in-group vs. out-group) and in-group favoritism. In other words, in-group
influence or in-group favoritism was observed only among those with high in-group
identification. However, when viewers scored low on in-group identification, in-group
favoritism disappeared. This finding is consistent with self-categorization theory; once people
identify themselves with their in-group, they become susceptible to its influence.
Second, the influence of the two moderators was contingent upon in-group
identification. The two factors—prototypical expertise of categorized groups and topic
35

involvement—moderated the influence of in-group UGC on viewers’ attitudes, but only for
those viewers with high in-group identification. Together, these findings suggest that ingroup identification is crucial to understand in-group influence on participatory websites.
Only those with heightened in-group identification utilized categorical cues to sort out
information which made certain users’ comments more influential than others.
When it comes to the findings from post hoc analyses, the question remains as to
how the valence of comments affects one’s evaluation of the source’s credibility. The finding
that commenters who left positive comments were perceived to be more credible than
commenters who left negative comments appears to be related to transfer of attitude
recursively (TAR) effect (Gawronski & Walther, 2008). Transfer of attitude recursively
effect suggests that evaluations endorsed by a source can be transferred to the source. For
instance, when a source evaluates a target in a negative manner, the source can acquire
negative evaluations. In Gawronski and Walther’s experiments (2008), pictures of several
sources were presented with evaluative statements (positive vs. negative). Specifically, four
sources evaluated their colleagues (targets) in positive ways and four sources evaluated their
colleagues in negative ways. The results indicate that a statement’s valence generated by a
source can be transferred back onto the source (sources leaving positive statements were
perceived to be more positive than sources leaving negative statements.) In this sense, the
current results appear to be consistent with TAR effects. Positive commenters were assessed
in more positive ways than negative commenters. However, even if the TAR effects were
observed across experimental conditions, since the effects of valence were only identified as
post hoc explanations, the effects remain empirically unverified.
Issues for Future Research
There are several issues that deserve attention in future research. First, the current
study employed the metacontrast principle to distinguish categorized groups (male vs.
36

female). Its employment, however, could affect the external validity of the research. On
participatory websites, multiple categorical cues are often accessible to viewers and thus the
distinction between an in-group and an out-group may not be as clear as depicted in the
current study. Previous literature indicates that multiple categorization criteria may hinder
categorized perceptions (in-group vs. out-group; Hall & Crisp, 2005). Crisp, Hewstone, and
Rubin (2001) argue that when individuals face multiple categorization criteria, in-group vs.
out-group distinctions become too complex to allow the sorting out of inconsistent
information, failing thus to trigger categorized perceptions. On the other hand, other studies
point out that when one category becomes relatively salient than others, it triggers
categorized perceptions. For instance, priming one of several alternative categories leads to
its activation and inhibition of alternative categories, which sparks categorized perceptions
(Macrae, Bodenhausen, & Milne, 1995). Another study argues that, even in the presence of
other categories, one category may become salient based on what other categories are present.
van Rijswijk and Ellemers (2002) point out that whether category-based perceptions are
triggered or not is determined by a comparative context (i.e., to which one category is mainly
compared). Specifically, the study argues that when evaluating targets that vary in multiple
categories, individuals tend to focus on a category that readily compares the targets and come
up with category-based perceptions using the category employed. These findings indicate that
the accessibility of multiple categories complicates the way categorical cues affect viewers’
attitudes. In order to extend our understanding of how categorical cues actually affect viewers’
attitudes on participatory websites, future research needs to examine 1) how categorical cues
affect viewers’ attitudes when multiple categorization criteria exist and 2) what makes certain
criteria more salient than others.
Second, the findings from the current study could be restricted due to the existence of
opposing messages (positive vs. negative UGC). For instance, this study employed the
37

metacontrast principle to distinguish groups. In order to distinguish groups (male vs. female),
participants viewed the two most positive and two most negative comments. However, it is
possible that the presence of a counter-argument could constrain the possible influence of
social categorical information. Meta analyses of one-sided vs. two-sided messages indicate
that two-sided messages without refutation lack persuasiveness compared to one-sided
persuasive messages or two-sided messages with refutation (Allen, 1991). Hale, Mongeau,
and Thomas (1991) suggest that arguments in two-sided, refutational messages lead to
positive evaluations and greater persuasive influence than do arguments in one-sided
messages or two-sided, non-refutational messages. In the present study, opposing arguments
were present, though none was refutational. Accordingly, the presence of the opposing
argument could have constrained the possible influence of social categorical cues embedded
in UGC.
Lastly, future study needs to synthesize how message-related and source-related
features of UGC affect viewers’ attitudes and perceptions along with the influence of
categorical cues. Post hoc analyses found that the credibility of the commenters was affected
by the valence of their comments. Even though evaluation of source failed to affect the
influence of in-group UGC, future research needs to expand our understanding of how people
actually utilize different kinds of information on participatory websites and how this leads to
social influence. For instance, the current study found that categorical cues trigger in-group
influence when individuals have high, rather than low, involvement. Future studies may
examine how, when involvement varies, in-group vs. out-group influence triggered by
categorical cues interacts with message-related (e.g., argument quality) and source-related
(e.g., source credibility) features of UGC. By synthesizing prior findings on UGC, future
studies may extend our understanding of how social influence occurs on participatory
websites.
38

Participatory websites display messages from multiple sources. The current study
focuses on how UGC affects viewers’ attitudes and perceptions. In particular, it captures the
dynamic nature of participatory websites by showing that categorical information about a
user that is embedded in UGC can activate in-group perception and in-group identification
among viewers. The findings imply that categorical information may allow viewers to feel
connected to certain users. Future research needs to further investigate how this dynamic
connection between a user and a viewer predicts social influence.

39

APPENDICES

40

Appendix A

Masculinity and femininity of a target business
Please indicate your impressions of this business.
Not masculine at all

123456789

Very masculine

Not professional at all

123456789

Very professional

Not appealing at all

123456789

Very appealing

Not female-oriented at all

123456789

Very female-oriented

Not favorable at all

123456789

Very favorable

Not feminine at all

123456789

Very feminine

Of poor quality

123456789

Of good quality

Not pleasant at all

123456789

Very pleasant

Not male-oriented at all

123456789

Very male-oriented

Valence
Please read each review and indicate how positive or negative each review sounds.
Strongly negative

1234567

Strongly positive

Helpfulness
Please indicate how helpful each review is in terms of making your decision.
Not helpful at all

1234567

Very helpful

Argument quality
Please indicate how strong of an argument each review presents. How would you evaluate
argument quality of each review?
Very weak

1234567

Masculinity and femininity
41

Very strong

Please indicate how masculine or feminine each review sounds.
Very masculine

1234567

Very feminine

Demographic questions
What is your sex?
What is your current class standing?
•
•
•
•
•

Freshman
Sophomore
Junior
Senior
Graduate student

What is your age?
What is your major?
How would you describe your primary ethnic heritage?
•
•
•
•
•
•
•

European
Asian
Hispanic
Pacific Islander
Native America
African
Others

If you have any comments or feedback about this study, please write them here.

42

Appendix B
Attitude toward a target object I
How would you rate your overall evaluation of the business?
Very negative

123456789

Very positive

How would you rate the quality of service this business offers?
Very negative

123456789

Very positive

Attitude toward a target object II (Holbrook & Batra, 1987; Mitchell & Olson, 1981) (9-point
semantic differential scale)
Dislike

123456789

Like

Negative

123456789

Positive

Favorable

123456789

Unfavorable

Of poor quality

123456789

Of good quality

Appealing

123456789

Unappealing

Bad

123456789

Good

Perceived expertise (adapted from Ohanian, 1990) (7-point Likert scale)
This reviewer seems to be an expert about the topic.
This reviewer seems to be experienced about the topic.
This reviewer seems to be knowledgeable about the topic.
This reviewer seems to be qualified to discuss the topic.
This reviewer seems to be skilled about the topic.

Perceived credibility (adapted from McCroskey and Teven, 1999) (7-point bipolar scale)
This reviewer seems to be…
Unintelligent

1234567

43

Intelligent

Untrained

1234567

Trained

Dishonest

1234567

Honest

Untrustworthy

1234567

Trustworthy

Not an expert

1234567

Expert

Dishonorable

1234567

Honorable

Immoral

1234567

Moral

Incompetent

1234567

Competent

Unethical

1234567

Ethical

Not understanding

1234567

Understanding

Prototypical expertise of a topic (adapted from Ohanian, 1990) (7-point Likert scale)
Men tend to be an expert about the topic.
Men tend to be experienced about the topic.
Men tend to be knowledgeable about the topic.
Men tend to be qualified to discuss the topic.
Men tend to be skilled about the topic.
Women tend to be an expert about the topic.
Women tend to be experienced about the topic.
Women tend to be knowledgeable about the topic.
Women tend to be qualified to discuss the topic.
Women tend to be skilled about the topic.

Social identification with in-group (adapted from Leach et al., 2008) (7-point Likert scale)
Being a woman/man is an important part of my self-image.
Being a woman/man is important to my sense of what kind of person I am.
Being a woman/man is an important reflection of who I am.
Being a woman/man has very little to do with how I feel about myself.
44

I have a lot in common with the average woman/man.
I am similar to the average woman/man.
I feel a bond with women/men.
Women/men have a lot in common with each other.
I identify with other women/men.
Being a woman/man just feels natural to me.
I feel personally implicated when women/men are criticized.

Masculinity and femininity of a target business
Please indicate your impressions of this business.
Not masculine at all

123456789

Very masculine

Not professional at all

123456789

Very professional

Not appealing at all

123456789

Very appealing

Not female-oriented at all

123456789

Very female-oriented

Not favorable at all

123456789

Very favorable

Not feminine at all

123456789

Very feminine

Of poor quality

123456789

Of good quality

Not pleasant at all

123456789

Very pleasant

Not male-oriented at all

123456789

Very male-oriented

Involvement (adapted from Cho & Boster, 2005) (7-point Likert scale)
Please read each statement and indicate the extent to which you agree or disagree.
The topic has little impact on my life.
My quality of life would not change depending on the topic.
The topic affects my life.
All in all, the effects of the topic on my life would be little.
It is easy for me to think how the topic influences my well-being.
45

Demographic questions
What is your sex?
What is your current class standing?
•
•
•
•
•

Freshman
Sophomore
Junior
Senior
Graduate student

What is your age?
What is your major?
How would you describe your primary ethnic heritage?
•
•
•
•
•
•
•

European
Asian
Hispanic
Pacific Islander
Native America
African
Others

If you have any comments or feedback about this study, please write them here.

46

Appendix C

Table 1 Sample size for experimental conditions
Topic
UGC Composition

Masculine

Feminine

Genderneutral

Male (+), Female (-)

n = 29

n = 31

n = 30

Male (-), Female (+)

n = 30

n = 29

n = 30

Male (+, -), Female (+, -)

n = 34

n = 31

n = 32

Note. (+) indicates positive comments and (-) indicates negative comments.

47

Table 2 Influence of the in-group commenters on viewers’ attitudes
In-group commenters’ position: positive
In-group commenters’ position: negative
In-group
A viewer’s
In-group
In-group
A viewer’s
In-group
position
position
influence
position
position
influence
9
1
1
1
1
9
9
2
2
1
2
8
9
3
3
1
3
7
9
4
4
1
4
6
9
5
5
1
5
5
9
6
6
1
6
4
9
7
7
1
7
3
9
8
8
1
8
2
9
9
9
1
9
1
Note. For ‘In-group position’ and ‘A viewer’s position’ columns, 1 = strongly negative, 9 =
strongly positive

Table 3 Influence of the out-group commenters on viewers’ attitudes
Out-group commenters’ position: positive
Out-group commenters’ position: negative
Out-group
A viewer’s
Out-group Out-group
A viewer’s
Out-group
position
position
influence
position
position
influence
9
1
1
1
1
9
9
2
2
1
2
8
9
3
3
1
3
7
9
4
4
1
4
6
9
5
5
1
5
5
9
6
6
1
6
4
9
7
7
1
7
3
9
8
8
1
8
2
9
9
9
1
9
1
Note. For ‘Out-group position’ and ‘A viewer’s position’ columns, 1 = strongly negative, 9 =
strongly positive

48

Table 4 Comment valence and credibility evaluation of commenters
Masculine topic
Male (+) Female (-)

Male (-) Female (+)

Credibility of M

M = 4.51, SD = 0.65, n = 29

M = 4.12, SD = 1.00, n = 31

Credibility of F

M = 4.05, SD = 0.82, n = 29

M = 4.15, SD = 0.86, n = 31

t (28) = 2.07, p = .02

t (30) = 0.16, p = .44

Comparison of M and F

Feminine topic
Male (+) Female (-)

Male (-) Female (+)

Credibility of M

M = 4.52, SD = 0.89, n = 30

M = 4.10, SD = 1.07, n = 29

Credibility of F

M = 4.10, SD = 0.92, n = 30

M = 4.77, SD = 0.69, n = 29

t (29) = 3.31, p = .002

t (28) = 3.36, p = .001

Comparison of M and F

Gender-neutral topic
Male (+) Female (-)

Male (-) Female (+)

Credibility of M

M = 4.56, SD = 0.57, n = 34

M = 4.30, SD = 0.65, n = 31

Credibility of F

M = 3.86, SD = 0.73, n = 34

M = 4.48, SD = 0.52, n = 31

t (33) = 4.49, p < .001

t (30) = 1.69, p = .05

Comparison of M and F

Note. M indicates male commenters and F indicates female commenters. (+) indicates
positive comments and (-) indicates negative comments.

49

REFERENCES

50

REFERENCES

Allen, M. (1991). Meta-analysis comparing the persuasiveness of one-sided and two-sided
messages. Western Journal of Speech Communication, 55, 390-404.
doi:10.1080/10570319109374395
Aronson, E., & Golden, B. W. (1962). The effect of relevant and irrelevant aspects of
communicator credibility on opinion change. Journal of Personality, 30, 135-146.
doi:10.1111/j.1467-6494.1962.tb01680.x
Arpan, L. M. (2002). When in Rome? The effects of spokesperson ethnicity on audience
evaluation of crisis communication. International Journal of Business
Communication, 39, 314-339. doi:10.1177/002194360203900302
Atkinson, D. R., Stephen, B., & Casas, J. M. (1981). Sexual preference similarity, attitude
similarity, and perceived counseling credibility and attractiveness. Journal of
Counseling Psychology, 28, 504-509. doi:10.1037/0022-0167.28.6.504
Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger
than good. Review of General Psychology, 5, 323-370. doi:10.1037/10892680.5.4.323
Carli, L. L. (2001). Gender and social influence. Journal of Social Issues, 57, 725-741.
doi:10.1111/0022-4537.00238
Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source
versus message cues in persuasion. Journal of Personality and Social Psychology,
39, 752-766. doi:10.1037/0022-3514.39.5.752
Cho, H. & Boster, F. J. (2005). Development and validation of value-, outcome-, and
impression-relevant involvement scales. Communication Research, 32, 235-264.
doi:10.1177/0093650204273764
Chu, S., & Kamal, S. (2008). The effect of perceived blogger credibility and argument quality
on message elaboration and brand attitudes. Journal of Interactive Advertising, 8,
26-37. doi:10.1080/15252019.2008.10722140
Cornetto, K, M. & Nowak, K. L. (2006). Utilizing usernames for sex categorization in
computer-mediated communication: Examining perceptions and accuracy.
Cyberpsychology and Behavior, 9, 377-387. doi:10.1089/cpb.2006.9.377
Diehl, M. (1990). The minimal group paradigm: Theoretical explanations and empirical
findings. European Review of Social Psychology, 1, 263-292.
doi:10.1080/14792779108401864
Edwards, C., Edwards, A., Qing, Q., & Wahl, S. (2007). The influence of computer-mediated

51

word-of-mouth communication on student perceptions of instructors and attitudes
toward learning course content. Communication Education, 53, 255-277.
doi:10.1080/03634520701236866
Fazio, R. H., & Dunton, B. C. (1997). Categorization by race: The impact of automatic and
controlled components of racial prejudice. Journal of Experimental Social
Psychology, 33, 451-470. doi:10.1006/jesp.1997.1330
Franck, K., & Rosen, E. (1949). A projective test of masculinity-femininity. Journal of
Consulting Psychology, 13, 247-256. doi:10.1037/h0057315
Ganster, T. (2011). You ‘like’, we ‘like’, I ‘like’? An empirical study about the impact of user
comments posted on commercial Facebook profiles with particular focus on
processes of social identification (Unpublished master’s thesis). University of
Duisburg-Essen, Germany.
Gawronski, B., & Walther, E. (2008). The TAR effect: When the ones who dislike become
the ones who are disliked. Personality and Social Psychology Bulletin, 34, 12761289. doi:10.1177/0146167208318952
Haiman, F. S. (1949). An experimental study of the effects of ethos in public speaking.
Speech Monograph, 16, 190-202. doi:10.1080/03637754909374974
Hains, S. C., Hogg, M. A., & Duck, J. M. (1997). Self-categorization and leadership: Effects
of group prototypicality and leader stereotypicality. Personality and Social
Psychology Bulletin, 23, 1087-1099. doi:10.1177/01461672972310009
Hale, J. L., Mongeau, P. A., & Thomas, R. M. (1991). Cognitive processing of one- and twosided persuasive messages. Western Journal of Speech Communication, 55, 380-389.
doi:10.1080/10570319109374394
Hall, N. R., & Crisp, R. J. (2005). Considering multiple criteria for social categorization can
reduce intergroup bias. Personality and Social Psychology Bulletin, 31, 1435-1444.
doi:10.1177/0146167205276084
Hamilton, M. A., & Nowak, K. L. (2010). Advancing a model of avatar evaluation and
selection. Psychnology Journal, 8, 33-65. Retrieved from
http://psychnology.org/File/PNJ8(1)/PSYCHNOLOGY_JOURNAL_8_1_HAMILT
ON.pdf
Heilburn, A. B. (1976). Measurement of masculine and feminine sex role identities as
independent dimensions. Journal of Consulting and Clinical Psychology, 44, 183190. doi:10.1037/h0076857
Hewstone, M., Hantzi, A., & Johnston, L. (1991). Social categorization and person memory:
The pervasiveness of race as an organizing principle. European Journal of Social
Psychology, 21, 517-528. doi:10.1002/ejsp.2420210606

52

Hicks, A., Comp, S., Horovitz, J., Hovarter, M., Miki, M., & Bevan, J. L. (2012). Why
people use Yelp.com: An exploration of uses and gratifications. Computers in
Human Behavior, 28, 2274-2279. doi:10.1016/j.chb.2012.06.034
Hogg, M. A. (2000). Subjective uncertainty reduction through self-categorization: A
motivational theory of social identity processes. European Review of Social
Psychology, 11, 223-255. doi:10.1080/14792772043000040
Hogg, M. A., & Terry, D. J. (2000). Social identity and self-categorization processes in
organizational contexts. Academy of Management Review, 25, 121-140.
doi:10.5465/AMR.2000.2791606
Hogg, M. A., & Tindale, R. S. (2005). Social identity, influence, and communication in small
groups. In J. Harwood & H. Giles (Eds.), Intergroup communication: Multiple
perspectives (pp.141-164). New York, NY: Peter Lang.
Hogg, M. A., Turner, J. C., & Davidson, B. (1990). Polarized norms and social frames of
reference: A test of the self-categorization theory of group polarization. Basic and
Applied Social Psychology, 11, 77-100. doi:10.1207/s15324834basp1101_6
Holbrook, M. B., & Batra, R. (1987). Assessing the role of emotions as mediators of
consumer responses to advertising. Journal of Consumer Research, 14, 404-420.
doi:10.1086/209123
Hollingshead, A, B., & Fraidin, S. N. (2003). Gender stereotypes and assumptions about
expertise in transactive memory. Journal of Experimental Social Psychology, 39,
355-363. doi:10.1016/S0022-1031(02)00549-8
Hornsey, M. J. (2008). Social identity theory and self-categorization theory: A historical
review. Social and Personality Psychology Compass, 21, 204-222.
doi:10.1111/j.1751-9004.2007.00066.x
Huddy, L., & Terkildsen, N. (1993). Gender stereotypes and the perception of male and
female candidates. American Journal of Political Science, 37, 119-147.
doi:10.2307/2111526
Jang, J.-W., & Walther, J. B. (2013). Social influence of user-generated content on
participatory websites. (Unpublished manuscript). Department of Communication,
Michigan State University, Michigan, USA.
Johnson, B. T., & Eagly, A. H. (1989). Effects of involvement on persuasion: A metaanalysis. Psychological Bulletin, 106, 290-314. doi:10.1037/0033-2909.106.2.290
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk.
Econometrica: Journal of the Econometric Society, 4. 362-377. Retrieved from
http://m.shookrun.com/documents/katvprospect.pdf

53

Karasawa, M. (1991). Toward an assessment of social identity: The structure of group
identification and its effects on in-group evaluations. British Journal of Social
Psychology, 30, 293-307. doi:10.1111/j.2044-8309.1991.tb00947.x
Krosnick, J. A., & Alwin, D. F. (1987). An evaluation of a cognitive theory of response-order
effects in survey measurement. Public Opinion Quarterly, 51, 201-219.
doi:10.1086/269029
Kurzban, R., Tooby, J., & Cosmides, L. (2001). Can race be erased? Coalitional computation
and social categorization. Proceedings of the National Academy of Sciences of the
United States of America, 98, 15387-15392. doi:10.1073/pnas.251541498
Laurent, G., & Kapferer, J.-N. (1985). Measuring consumer involvement profiles. Journal of
Marketing Research, 22, 41-53. doi:10.2307/3151549
Leach, C. W., van Zomeren, M., Zebel, S., Vliek, M. W., Pennekamp, S. F., Doosje, B.,
Ouwerkerk, J. W., Spears, R. (2008). Group-level self-definition and self-investment.
Journal of Personality and Social Psychology, 95, 144-165. doi:10.1037/00223514.95.1.144
Lee, E.-J. (2007). Categorical person perception in computer-mediated communication:
Effects of character representation and knowledge bias on sex inference and
informational social influence. Media Psychology, 9, 309-329.
doi:10.1080/15213260701286007
Lee, E.-J. (2012). That’s not the way it is: How user-generated comments on the news affect
perceived media bias. Journal of Computer-mediated Communication, 8, 32-45.
doi:10.1111/j.1083-6101.2012.01597.x
Lee, E.-J., Jang, J.-W., & Kim, M. (2009, November). Interpersonal interactivity in online
journalism: What do readers’ comments on internet news sites tell us? Paper
presented at National Communication Association Annual Conference, Chicago, IL.
Lee, E.-J., & Jang, Y. J. (2010). What do others’ reactions to news on Internet portal
sites tell us? Effects of presentation format and readers’ need for cognition on reality
perception. Communication Research, 37, 825-846. doi:10.1177/0093650210376189
Lee, H., Herr, P. M., Kardes, F. R., & Kim, C. (1999). Motivated search: Effects of choice
accountability, issue involvement, and prior knowledge on information acquisition
and use. Journal of Business Research, 45, 75-88. doi:10.1016/S01482963(98)00067-8
Lee, J., Park, D.-H., & Han, I. (2008). The effect of negative online consumer reviews on
product attitude: An information processing view. Electronic Commerce Research
and Application, 7, 341-352. doi:10.1016/j.elerap.2007.05.00
Lee, M., Rodgers, S., & Kim, M. (2009). Effects of valence and extremity of eWOM on
attitude toward the brand and website. Journal of Current Issues and Research in
Advertising, 31, 1-12. doi:10.1080/10641734.2009.10505262
54

Luo, C., Lan, Y., Wang, C., & Ma, L. (2013, June). The effect of information consistency and
information aggregation on eWOM readers’ perception of information overload.
Paper presented at Pacific Asia Conference on Information System, Korea.
Mackie, D. M. (1986). Social identification effects in group polarization. Journal of
Personality and Social Psychology, 50, 720-728. doi:10.1037/0022-3514.50.4.720
Mackie, D. M., Worth, L. T., & Asuncion, A. G. (1990). Processing of persuasive in-group
messages. Journal of Personality and Social Psychology, 58, 812-822.
doi:10.1037/0022-3514.58.5.812
Maheswaran, D., & Meyers-Levy, J. (1990). The influence of message framing and issue
involvement. Journal of Marketing Research, 27, 361-367. doi:10.2307/3172593
Malhotra, N. K. (1984). Reflections on the information overload paradigm in consumer
decision making. Journal of Consumer Research, 10, 436-440. doi:10.1086/208982
McCroskey, J. C., & Teven, J. J. (1999). Good will: A reexamination of the construct and its
measurement. Communication Monograph, 66, 90-103.
doi:10.1080/03637759909376464
Mitchell, A. A., & Olson, J. C. (1981). Are product attribute beliefs the only mediator of
advertising effects on brand attitude? Journal of Marketing Research, 18, 318-332.
doi:10.2307/3150973
Mullin, B.-A., & Hogg, M. A. (1999). Motivations for group membership: The role of
subjective importance and uncertainty reduction. Basic and Applied Social
Psychology, 21, 91-102. doi:10.1207/S15324834BA210202
Oakes, P. J., Turner, J. C., & Haslam, S. A. (1991). Perceiving people as group members: The
role of fit in the salience of social categorizations. British Journal of Social
Psychology, 30, 125-144. doi:10.1111/j.2044-8309.1991.tb00930.x
Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers’
perceived expertise, trustworthiness, and attractiveness. Journal of Advertising, 19,
39-52. doi:10.1080/00913367.1990.10673191
O’Keefe, D. J. (2002). Persuasion: Theory and research. Los Angeles, CA: Sage.
Park, D. H., & J. Lee. (2008). Electronic word-of-mouth overload and its effect on consumer
behavioral intention depending on consumer involvement. Electronic Commerce
Research and Applications, 7, 386-398. doi:10.1016/j.elerap.2007.11.004
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. doi:10.2753/JEC1086-4415110405

55

Petty, R. E., & Cacioppo, J. T. (1979). Issue involvement can increase or decrease persuasion
by enhancing messages-relevant cognitive responses. Journal of Personality and
Social Psychology, 37, 1915-1926. doi:10.1037/0022-3514.37.10.1915
Pornpitakpan, C. (2004). The persuasiveness of source credibility: A critical review of five
decades’ evidence. Journal of Applied Social Psychology, 34, 243-281.
doi:10.1111/j.1559-1816.2004.tb02547.x
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing
and comparing indirect effects in multiple mediators models. Behavior Research
Methods, 40, 879-891. doi:10.3758/BRM.40.3.879
Racherla, P., & Friske, W. (2012). Perceived ‘usefulness’ of online consumer reviews: An
exploratory investigation across three services categories. Electronic Commerce
Research and Applications. 11, 548-559. doi:10.1016/j.elerap.2012.06.003
Reid, S. A., Giles, H. & Harwood, J. (2005). A self-categorization perspective on
communication and intergroup relations. In J. Harwood & H. Giles (Eds.),
Intergroup communication: Multiple perspectives (pp.241-263). New York, NY:
Peter Lang.
Reid, S. A., & Hogg, M. A. (2005). A self-categorization explanation for the third-person
effect. Human Communication Research, 31, 129-161. doi:10.1111/j.14682958.2005.tb00867.x
Ren, Y., Harper, F. M., Drenner, S., Terveen, L., Kiesler, S., Riedl, J., & Kraut, R. E. (2012).
Building member attachment in online communities: Applying theories of group
identity and interpersonal bonds. MIS Quarterly, 36, 841-864. Retrieved from
http://www.cs.cmu.edu/~kiesler/publications/2012/building-member-attachmentgroup-identify.pdf
Rozin, P., & Royzman, E. B. (2001). Negativity bias, negativity dominance, and contagion.
Personality and Social Psychology Review, 5, 296-320.
doi:10.1207/S15327957PSPR0504_2
Schmader, T. (2002). Gender identification moderates stereotypes threat effects on women’s
math performance. Journal of Experimental Social Psychology, 38, 194-201.
doi:10.1006/jesp.2001.1500
Simon, B., Hastedt, C., & Aufderheide, B. (1997). When self-categorization makes sense:
The role of meaningful social categorization in minority and majority members’
self-perception. Journal of Personality and Social Psychology, 73, 2, 310-320.
doi:10.1037/0022-3514.73.2.310
Simons, H. W., Berkowitz, N. N., & Moyer, R. J. (1970). Similarity, credibility, and attitude
change: A review and a theory. Psychological Bulletin, 73, 1-16.
doi:10.1037/h0028429

56

Spencer, S. J., Steele, C. M., & Quinn, D. M. (1999). Stereotype threat and women’s math
performance. Journal of Experimental Social Psychology, 35, 4-28.
doi:10.1006/jesp.1998.1373
Spottswood, E. L., Walther, J. B., Holmstrom, A. R., & Ellison, N. E. (2013). Personcentered emotional support and gender attributions in computer-mediated
communication. Human Communication Research, 39, 295-316.
doi:10.1111/hcre.12006
Stangor, C., Lynch, L., Duan, C., & Glass, B. (1992). Categorization of in individuals on the
basis of multiple social features. Journal of Personality and Social Psychology, 62,
207-218. doi:10.1037/0022-3514.62.2.207
Stone, J., Lynch, C. I., Sjomeling, M., & Darley, J. M. (1999). Stereotype threat effects on
Black and White athletic performance. Journal of Personality and Social
Psychology, 77, 1213-1227. doi:10.1037/0022-3514.77.6.1213
Turner, J. C. (1987). Rediscovering the social group: A self-categorization theory. New York,
NY: Basil Blackwell.
Turner, J. C. (1991). Social influence. Pacific Grove, CA: Brooks Cole.
Turner, J. C., & Oakes, P. J. (1989). Self-categorization theory and social influence. In P. B.
Paulus (Ed.), The psychology of group influence (pp.355-373). Hillsdale, NJ:
Erlbaum.
van Rijswijk, W., & Ellemers, N. (2002). Context effects on the application of stereotype
content to multiple categorizable targets. Personality and Social Psychology Bulletin,
28, 90-101. doi:10.1177/0146167202281008
Voci, A. (2006). Relevance of social categories, depersonalization and group processes: Two
field tests of self-categorization theory. European Journal of Social Psychology, 36,
73-90. doi:10.1002/ejsp.259
Walther, J. B., DeAndrea, D., Kim, J., & Anthony, J. C. (2010). The influence of online
comments on perceptions of antimarijuana public service announcements on
YouTube. Human Communication Research, 36, 469-492. doi:10.1111/j.14682958.2010.01384.x
Walther, J. B., & Jang, J.-W. (2012). Communication processes in participatory websites.
Journal of Computer-mediated Communication, 18, 2-15. doi:10.1111/j.10836101.2012.01592.x
Walther, J. B., Liang, Y., Ganster, T., Wohn, D. Y., & Emington, J. (2012). Online reviews,
helpfulness ratings, and consumer attitudes: An extension of congruity theory to
multiple sources in Web 2.0. Journal of Computer-mediated Communication, 18,
97-112. doi:10.1111/j.1083-6101.2012.01595.x

57

Weaver, D. H. (1980). Audience need for orientation and media effects. Communication
Research, 7, 361-373. doi:10.1177/009365028000700305
Willemsen, L. M., Neijens, P. C., & Bronner, F. (2012). The ironic effect of source
identification on the perceived credibility of online product reviewers. Journal of
Computer-mediated Communication, 18, 16-31. doi:10.1111/j.10836101.2012.01598.x
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, 101-112. doi:10.1007/BF02894421
Winter, S., & Krämer, N. C. (2012). Selecting science information in web 2.0: How source
cues, message sidedness, and need for cognition influence users’ exposure to blog
posts. Journal of Computer-mediated Communication, 18, 80-96.
doi:10.1111/j.1083-6101.2012.01596.x
Yu, J., & Cooper, H. (1983). A quantative review of research design effects on response rates
to questionnaires. Journal of Marketing Research, 20, 36-44. doi:10.2307/3151410

58