Sunteți pe pagina 1din 20

Human Communication Research ISSN 0360-3989

ORIGINAL ARTICLE

Leave Your Comment Below: Can Biased


Online Comments Inuence Our Own
Prejudicial Attitudes and Behaviors?
Mark Hsueh1 , Kumar Yogeeswaran1 , & Sanna Malinen2
1 Department of Psychology, University of Canterbury, Christchurch 8041, New Zealand
2 Department of Management, University of Canterbury, Christchurch 8041, New Zealand

Increased use of online communication in everyday life presents a growing need to under-
stand how people are influenced by others in such settings. In this study, online comments
established social norms that directly influenced readers expressions of prejudice both con-
sciously and unconsciously. Participants read an online article and were then exposed to
antiprejudiced or prejudiced comments allegedly posted by other users. Results revealed
that exposure to prejudiced (relative to antiprejudiced) comments influenced respondents
to post more prejudiced comments themselves. In addition, these effects generalized to par-
ticipants unconscious and conscious attitudes toward the target group once offline. These
findings suggest that simple exposure to social information can impact our attitudes and
behavior, suggesting potential avenues for social change in online environments.

Keywords: Computer-Mediated Communication, Social Norms, Implicit Prejudice, Explicit


Prejudice, Online Comments.

doi:10.1111/hcre.12059

Computer-mediated communications (CMC; such as online forums, blogs, Facebook,


and Twitter) have become routine ways of communicating, with over 500 million
tweets shared per day (Internet Live Stats, 2014). Worldwide, people are also turn-
ing to online news sites in record numbers to obtain information on current affairs
(Jurkowitz, 2014). What was once a one-to-many asynchronous channel has evolved
into an interactive medium where readers are also contributors, in what is referred
to as Web 2.0 (Walther, DeAndrea, Kim, & Anthony, 2010; Walther & Jang, 2012).
Although less personal due to the lack of visual cues, modern social media reaches
a broader audience almost immediately and can appear more credible because of the
messages appearing in (digital) print (Jansen, Zhang, Sobel, & Chowdury, 2009). In
Web 2.0, user comments can have more influence over readers opinions than the con-
tent proprietor (Walther et al., 2010), and the sharing of information between online

Corresponding author: Kumar Yogeeswaran; e-mail: kumar.yogeeswaran@canterbury.ac.nz

Human Communication Research 41 (2015) 557576 2015 International Communication Association 557
Online Comments on Prejudice Expression M. Hsueh et al.

groups of like-minded individuals can even form powerful efforts such as the 2010
Egyptian uprising (Walther & Jang, 2012).
As readers have evolved to become content creators in Web 2.0, recipients find it
difficult to determine what information is credible and what is not (Metzger, Flanagin,
& Medders, 2010). For example, Twitter is commonly used by customers to ascertain
the quality of a companys product, and this microblogging can have an incredible
impact on a corporations public image (Jansen et al., 2009). However, the monitoring
of Web-based content is largely absent, which may result in readers making decisions
based on erroneous or biased information (Metzger et al., 2010).
Further impeding the growth of constructive online environments are com-
menters who express uncivil, hateful, and prejudiced comments they would be
unlikely to express in face-to-face settings (Anderson, Brossard, Scheufele, Xenos,
& Ladwig, 2013). In fact, many Websites and corporations struggle to deal with
antisocial and prejudiced behaviors from users (Brisbane, 2010; Houston, Hansen,
& Nisbett, 2011; Jansen et al., 2009; Popular Science, 2013), such as the recent U.S.
Coca-Cola advertisement celebrating diversity that was met with hostile reactions
from some viewers (Lee, 2014). Dissenting comments can detract and reduce the
importance of contemporary social issues (Thorson, Vraga, & Ekdale, 2010), and
even public service announcements (PSA) promoting antidrug use can be met with
derisive attitudes (Walther et al., 2010).
To combat this, some sites are encouraging users to sign in with personally
identifying information (e.g., YouTube); hiring community managers to moderate
comments (Anderson et al., 2013); or choosing to completely dismantle their reader
comment sections (Hughey & Daniels, 2013; Popular Science, 2013). However,
research has not examined whether these prejudiced online sentiments actually affect
readers own personal attitudes.
This research examines the following questions: Do online comments posted by
other users impact an individuals own expressions of prejudice? Specifically, can such
social influence impact ones own comments as well as their conscious and uncon-
scious prejudicial attitudes? Considering the foundation of social interactions in Web
2.0, this research incorporates communication literature on CMC and social psychol-
ogy research on social norms, prejudice expression, and implicit social cognition to
address these questions.

The influence of social norms on prejudice expression


Research in communication and social psychology on norms are intertwined as
social norms are created and spread via communication avenues (Lapinski & Rimal,
2005; Yanovitzky & Rimal, 2006), such as online media (Hogg & Reid, 2006). Various
studies in social psychology have found that group norms can be a key component
in predicting behavioral tendencies (Crandall, Eshleman, & OBrien, 2002; Hogg &
Reid, 2006; Jansen et al., 2009; Lee, 2006; Stangor, Sechrist, & Jost, 2001; Yanovitzky
& Rimal, 2006). Communication and social theorists argue that not fitting in with
the social norm may result in individuals feeling socially ostracized (Hogg & Reid,

558 Human Communication Research 41 (2015) 557576 2015 International Communication Association
M. Hsueh et al. Online Comments on Prejudice Expression

2006; Lapinski & Rimal, 2005; Rimal & Real, 2003), while conforming to the salient
group norm is beneficial to both the individual and the group (Cialdini & Goldstein,
2004).
Social norms indicate that individuals look to others in their group for a socially
acceptable response and then mimic the behaviors and attitudes of the groups salient
norm (Blanchard, Crandall, Brigham, & Vaughn, 1994; Blanchard, Lilly, & Vaughn,
1991; Deutsch & Gerard, 1955; Lapinski & Rimal, 2005; Zitek & Hebl, 2007). Further,
people show increased intergroup affection in groups that focus on egalitarian values
(Blanchard et al., 1991; Schultz, Nolan, Cialdini, Goldstein, & Griskevicius, 2007), and
biased intergroup attitudes while in prejudiced groups (Blanchard et al., 1994).
This indicates that individual prejudicial attitudes may actually be a reflection
of the attitudes of their immediate group norms (Lapinski & Rimal, 2005; Plant &
Devine, 1998). For example, Blanchard et al. (1994) had a confederate voice either
prejudiced or antiprejudiced opinions when researchers asked participants to rate
Black Americans. Their data revealed that antiracist messages from the confederate
significantly heightened antiracist sentiments on a college campus relative to a con-
trol, while a prejudiced message from the confederate significantly reduced antiracist
sentiments relative to controls. Similarly, Stangor et al. (2001) found that they could
influence prejudicial attitudes of university students toward Black student peers by
informing participants how other fictitious students responded on surveys.
These findings suggest that perceiving a social norm as supporting or opposing
bigotry can influence an individuals own responses in the given context. However,
little research has addressed the influence of group norms on prejudice expression in
an online CMC environment. Moreover, the question remains whether group norms
continue to exert influence on the individual even after they have left the specific con-
text. This study aims to address these gaps by applying psychological theory from
prejudice expression and social norms to the field of online communication. Specif-
ically, if people in an online context express prejudiced (relative to antiprejudiced)
sentiments, would readers look upon these sentiments as the socially accepted norm
and express more prejudiced sentiments themselves?

Online social influence


Research on persuasion in online contexts has found that users actively seek the
opinions of others to inform opinions and, in particular, consumer decisions (Ardelet
& Briar, 2011; Cabezudo, Izquierdo, & Pinto, 2013; Metzger et al., 2010). As CMC
environments often lack the physical aspects of traditional face-to-face interactions,
such as facial expressions and vocal fluctuations (Woong-Yun & Park, 2011), users
learn to conduct themselves by taking cues from any available online interaction
(McLaughlin & Vitak, 2011; Walther & Parks, 2002). The CMC environment makes
people more susceptible to these group norm effects due to a combination of deper-
sonalization and group immersion, where people tend to perceive themselves and
others as social groups rather than individuals (Postmes, Spears, & Lea, 2000; Reicher,
Spears, & Postmes, 1995).

Human Communication Research 41 (2015) 557576 2015 International Communication Association 559
Online Comments on Prejudice Expression M. Hsueh et al.

Studies have found that these CMC groups tend to form their own identities and
characteristics (Postmes & Spears, 2002; Spears, Lea, Corneliussen, Postmes, & Ter
Harr, 2002), which can have even stronger group identities and norms than normal
face-to-face groups because of the interchangeability of its group members (Lee, 2006;
Postmes et al., 2000; Spears et al., 2002). This effect is argued to be even stronger
when individuals are isolated due to the enhancement or undermining of status effects
among in-group members and other out-groups (Spears et al., 2002).
The behaviors toward out-groups are significantly different than how one speaks
to their in-group, and when changing groups the salient norm/identity also changes,
indicating that group norms override individualism (Postmes et al., 2000). This sug-
gests that by setting a group norm to behave a certain way, new members will alter
their identity to fit with the salient group norm. However, this can also lead to the
spreading of inaccurate information as readers rarely or only occasionally verify
information presented to them by other users (Metzger et al., 2010).
All comments (prejudiced and antiprejudiced) can potentially impact readers as
user comments are used by observers to gauge public opinions (Walther & Jang, 2012),
which can result in an information bias (Houston et al., 2011). In a study by Edwards,
Edwards, Qing, and Wahl (2007), students were presented with either negative or pos-
itive comments about a professor; then after watching a clip of that professor teaching,
participants were asked to rate the professor online. Results revealed that the alleged
comments affected participants ratings of the professors attractiveness, credibility,
and their personal attitudes and motivation toward the course. This further suggests
that social information, regardless of its accuracy, plays a key part in attitude devel-
opment (Majchrzak, Faraj, Kane, & Azad, 2013; Metzger et al., 2010).
However, in a Web 2.0 setting, users are not required to adhere to the same social
responsibilities that mainstream Websites do (Dylko & McCluskey, 2012) and, iron-
ically, the more extreme the message bias, the greater the social influence (Walther
& Jang, 2012). These findings are consistent with the literature on persuasion which
posits that individuals rely on social network evaluations of information to develop
their views (Hughes et al., 2014; Lapinski & Rimal, 2005), including the use of others
opinions to confirm ones own view (Metzger et al., 2010). We therefore hypothesize
that other users prejudiced comments can influence the prejudicial attitudes of sub-
sequent readers.

Implicit and explicit prejudice expressions in online comments


Although prejudice expression in many societies around the world has declined over
the last half-century, it is difficult to determine whether this change reflects increased
social consciousness, or largely egalitarian shifts in societal norms that discourage
people from expressing blatant forms of prejudice (Crandall et al., 2002; Toosi, Bab-
bitt, Ambady, & Sommers, 2012). Although people base their comments on their
respective group norms, the expressed attitudes of individuals are not necessarily the
same as their underlying attitudes (Hogg & Reid, 2006). As self-reported conscious
(explicit) measures are prone to social desirability biases (Greenwald, Poehlmann,

560 Human Communication Research 41 (2015) 557576 2015 International Communication Association
M. Hsueh et al. Online Comments on Prejudice Expression

Uhlmann, & Banaji, 2009; Sinclair, Lowery, Hardin, & Colangelo, 2005), ascertain-
ing whether these egalitarian social norms are actually being internalized on a deeper
subconscious level or if they have just been adopted at face value to fit in with greater
societal norms is difficult (Greenwald et al., 2009).
To help alleviate this problem, new tools to assess subtle or unconscious (implicit)
attitudes and beliefs about others were created (Dasgupta, 2004; Fazio & Olson, 2003;
Greenwald, McGhee, & Schwartz, 1998; Nosek et al., 2007). One such measure of
implicit prejudice is the implicit association test (IAT; Greenwald et al., 1998), which
has been found to be relatively immune to the effects of individual self-presentation.
In a meta-analysis, Greenwald et al. (2009) demonstrated that the IAT was able to
better predict behavior and judgments than explicit self-report measures, especially
in socially sensitive contexts such as intergroup bias. These new measures offer insight
into prejudicial attitudes that emerge automatically without ones conscious awareness
(Dasgupta, 2004; Fazio & Olson, 2003), with various studies revealing that implicit
attitudes and stereotypes can predict discriminatory behaviors and judgments in job
hiring, nonverbal interactions, voting, medical decisions, and evaluations of public
policy (for a review, see Greenwald et al., 2009).
Although automatic and beyond conscious awareness, implicit attitudes, and
stereotypes are not fixed, as established by over a decade of research (for a review,
see Blair, 2002; Dasgupta, 2009). These implicit attitudes and stereotypes are mal-
leable based on social context, exposure to counterstereotypes, and social relations
among other variables (Blair, 2002; Dasgupta, 2009). Although reading prejudiced
or antiprejudiced comments may intuitively be expected to shape explicit prejudice,
whether these can also become internalized to influence automatic or unconscious
prejudice given its subtle nature remains to be seen. Therefore, given the growing
importance of implicit attitudes and stereotypes, this research will examine the
impact of online social norms on readers subsequent implicit attitudes as well as
self-reported, explicit prejudice, toward an entire ethnic group. In this research, we
hypothesize the following:
H1: Participants exposed to prejudiced social norms established by online comments are
expected to post more prejudiced online comments themselves relative to participants
exposed to antiprejudiced social norms via their online comments.
H2: Participants exposed to prejudiced online comments will express more explicit
prejudice compared with participants exposed to antiprejudiced online comments who
will respond with more antiprejudiced sentiments.
H3: Participants exposed to prejudiced online comments will show more implicit
prejudice relative to participants exposed to antiprejudiced online comments.

Method
Participants
A total of 137 volunteers in New Zealand took part in the study (98 females, 37 males,
2 undisclosed) in exchange for a $10 gift voucher or course credit. These participants
were recruited through advertisement/flyers posted around a university campus or

Human Communication Research 41 (2015) 557576 2015 International Communication Association 561
Online Comments on Prejudice Expression M. Hsueh et al.

from an introductory psychology participant pool. Ages of participants ranged from


18 to 50 (M = 22.79, SD = 6.68), whereas the ethnic backgrounds of the participants
were fairly typical of the local population with participants indicating European (95),
Asian (21), Maori (8), Pacific Islanders (4), Indian (3), or other (6) ethnicities.

Materials
News article
Participants were presented with a Marketing Survey Website, which was created
specifically for this study using the Website creator within www.weebly.com. To avoid
suspicion about the authenticity of the Website, the domain name www.digitalpoll.org
was created for the alleged research Website. The marketing research Website
explained that the Tertiary Education Commission (TEC) was planning to increase
scholarship funding support for international students, with a heavy emphasis toward
students from East Asia. However, due to recent claims of Chinese-speaking students
being caught cheating in their tertiary studies, the public perception of the proposal
is currently uncertain.
The content for this article was adapted directly from recent news stories published
in major domestic newspapers. Readers were then told that because of the uncertainty
in the new program, the TEC had hired an external research firm to gather feedback
from universities across the country to ascertain the current stance of public opinion
toward the proposed funding increase. We specifically chose such a news story as it
directly relates to university students and education, meaning that participants would
be more invested in the issue.
After reading the TEC article, participants were asked to offer their own com-
ments. However, before they could place their comments, participants had to scroll
past comments placed by the experimenter masking as other participants. By setting
up the experiment in this way, it maintained ecological validity akin to the normal
process that Internet users would go through before entering their own comments.
Although it was not explicitly stated, the participants were free to move onto the next
step without leaving a comment. The news article remained the same across both con-
ditions with only the comments that followed the article differing.

Social norm manipulation


Taking cues from the Thorson et al. (2010) paradigm, the social norm environment
was created using 24 preselected comments (12 comments in the Antiprejudice
norm; 12 comments in the Prejudice norm). Comments were re-entered every
fortnight to keep the time stamps recent. All comments were copied from actual
responses to a recent news article about a ghostwriting firm writing assignments for
Chinese-speaking students (Stuff, 2013). User names of the commenters were all fic-
tional and included both real sounding names that appeared mostly White European
(e.g., charlescarter, martin35) or code-like names (e.g., GenX, badabing).
A spectrum of comments was chosen to present differing levels of antiprejudiced
or prejudiced expression to maintain ecological validity. In the antiprejudice condi-
tion, all 12 comments were supportive of Asian international students and cautioned

562 Human Communication Research 41 (2015) 557576 2015 International Communication Association
M. Hsueh et al. Online Comments on Prejudice Expression

people not to generalize negative feelings toward all Asian people (e.g., I would like
to take this opportunity to remind everyone reading the article to keep a level head and
refrain from making these comments targeted at all Chinese students. Most Chinese stu-
dents in fact are hardworking intelligent people who receive good grades from their own
efforts, If your trying to tell us all that its only Chinese that are cheating then you are
nave.).
By contrast, in the prejudice condition, all 12 comments expressed dissent toward
increasing support for Asian students and revealed suspicions about them (e.g., It
wasnt that long ago Asians not just Chinese immigrants that were getting their driver
license illegally also. As someone stated its the Chinese way. Then let it stay in China
and not here in NZ. Those involved should have their qualification stripped. At Lincoln
University 80% of students caught cheating are form mainland China. Fact.) These
comments were, however, not extremely hateful and did not call for any violent actions
toward any group as such comments were not placed in response to this particular
news piece. In addition, we wanted users to believe that these comments were placed
by their peers and not members of any radical hate group.

Measures
Explicit prejudice responses
Semantic differentials were used to assess explicit prejudice toward Asians similar
to previous research (Glick & Fiske, 1996; Hewstone, Islam, & Judd, 1993; Sagar
& Schofield, 1980). Two semantic differential items were included (Glick & Fiske,
1996; Hewstone et al., 1993; Ho & Jackson, 2001); How positive do you feel towards
Asians, and How much do you like Asians, which were rated on a 7-point scale,
with higher scores indicating more positive attitudes toward Asians. These items
were combined after ensuring that they show high internal consistency ( = .87). In
addition, a Feeling Thermometer was included, which asked participants to rate
on a 1100 scale how they felt toward Asians, with lower scores indicating cold or
unfavorable feelings, and higher scores indicating warm or favorable feelings. These
were examined separately as they have different anchors, but also combined in a
multivariate analysis of variance (MANOVA).

Implicit prejudice responses


Implicit prejudice toward Asians was assessed using a race IAT (Greenwald et al.,
1998), which uses response latency as an indicator of the strength of association
between a particular ethnic group and positive versus negative valence (for a
demonstration, visit https://implicit.harvard.edu/implicit/). Following the procedure
described by Greenwald, Nosek, and Banaji (2003), this IAT consisted of seven
blocks of trials in which participants are asked to categorize stimuli as quickly and
accurately as they can. In each block, a selection of six Asian faces (three males and
three females), six White faces (three males and three females), six positive words
(e.g., kindness), or six negative words (e.g., sickness) individually appeared in the
center of the screen. Whenever a participant responded incorrectly, an X message

Human Communication Research 41 (2015) 557576 2015 International Communication Association 563
Online Comments on Prejudice Expression M. Hsueh et al.

would appear to inform the participant that they responded incorrectly and were
then required to press the correct response key to advance. The time between the
stimuli appearing on the screen and the participants correct response was recorded
(for details, see Greenwald et al., 2003).
Participants first completed a single-category practice block where they only cat-
egorized good and bad words (20 trials) using a A or 5 (num-pad) key. They then
completed another single-category practice block where they categorized White and
Asian faces (20 trials) using the same two keys. Next, these two tasks were combined
such that participants would have to categorize all four types of stimuli simultane-
ously using the same two response keys in a short practice block (20 trials) and a
critical block (50 trials).
Half of the participants were instructed to use one key to categorize White faces
or Good words and the other key to categorize Asian faces or Bad words, while the
remaining participants were instructed to use one response key to categorize Asian
faces or Good words, and the other key for White faces or Bad words. Then, par-
ticipants completed another single-category practice block (20 trials) in which the
response key assignment for White and Asian faces was reversed (i.e., if they pressed
A for Asian faces and 5 for White faces before, they would now press 5 for Asian
faces and A for White faces).
After this, participants once again completed a practice block (20 trials) and crit-
ical block (50 trials), where they simultaneously categorized all four types of stimuli.
Those who previously classified White faces and Good words on one response key ver-
sus Asian faces and Bad words on the other key now completed the reverse pairings.
Differences in reaction times and the pooled standard deviation of response
times were later used to calculate an IAT D score which compares response times
of individuals against deviations from their own response times in order to create a
scale unit that reduces contamination from external factors. Positive IAT D scores
indicate stronger associations with White-Good/Asian-Bad relative to Asian-Good/
White-Bad (i.e., a more prejudiced response). Negative deviation scores indi-
cate stronger associations with Asian-Good/White-Bad relative to White-Good/
Asian-Bad associations (i.e., a less prejudiced response).

Design
The independent variable included the two social norm conditions, whereas the
dependent variables included prejudice expressions in the forms of comment ratings
as a response to the online article, an IAT D score, and a self-report prejudice
questionnaire.

Procedure
Participants were recruited under the guise of completing tasks from several different
studies because each of these studies were too short on their own to justify full pay-
ment or course credit. Upon entering the laboratory, participants were taken to one
room where they were provided with a paper questionnaire including a demographic

564 Human Communication Research 41 (2015) 557576 2015 International Communication Association
M. Hsueh et al. Online Comments on Prejudice Expression

survey. Upon completing this, they were taken to a different room containing four par-
titioned cubicles each with its own computer station. Each participant was seated at a
work station and introduced to the study, presented as an International Student Edu-
cation study that was being conducted by an external data collection agency on behalf
of the TEC. The participants were linked to the marketing research Website where
they read the news article described earlier. After reading the article, participants were
asked to offer their feedback about the government plan. To do so, participants were
required to scroll past each of the norm conditioning comments.
To parallel real life, we did not ask participants to read any particular number
of comments, but the format of the Website ensured that participants would have
to at least skim past others comments before placing their own. Using their name
or a random code,1 participants placed their own comment on the Website, which
was seen only on their own screen and not viewable to others in order to ensure that
each participant saw the exact same content. Participants were then told that they
would now complete a different set of tasks as part of another study. Under this guise,
they completed the IAT and the self-report questionnaire, the orders of which were
counterbalanced between subjects.

Results
A comparison of gender effects showed no interaction between gender and norm
condition on any of the DVs (p > .33). A factorial analysis of variance (ANOVA)
was conducted to examine interactive effects caused between Asian (n = 28) versus
non-Asian participants (n = 108). The results indicated no interaction effects across
any of the DVs (p > .16). Demographic factors are therefore not discussed in further
analyses.

Online comments
The online comments placed by the participants were graded on a 17 scale by two
independent coders who were both graduate students in psychology and blind to
each participants condition. These coders were independently asked to rate the com-
ments on a 17 scale regarding the extent to which they perceived them to be preju-
diced toward Asians (implying clear negative sentiments toward the group), neutral,
or antiprejudiced or pro-Asian sentiments (implying unprejudiced or protective feel-
ings about Asians). Specifically, these comments were coded such that 1 indicated
pro-Asian sentiments (e.g., I think it is really positive to be trying to boost the num-
ber of Chinese coming to study in New Zealand. It is helping our economy but also
pushing to establish positive relationships with the Chinese.); 4 indicated neutral or
nonrelated sentiments (e.g., I do not think that cheating is limited to any particu-
lar ethic group); and 7 indicated anti-Asian sentiments (e.g., Chinese people have
a bad rep for a reason). The two ratings were indexed by averaging the rater scores to
provide one composite of comment ratings because the average intraclass correlation
coefficient revealed a strong overlap between the ratings (ICC = .82). Two outliers (i.e.,

Human Communication Research 41 (2015) 557576 2015 International Communication Association 565
Online Comments on Prejudice Expression M. Hsueh et al.

Table 1 Descriptive Statistics and Significance of Independent Means t-tests for


Antiprejudice and Prejudice Social Norm Conditions for Each Dependent Variable

Antiprejudice norm Prejudice norm


Measure Range Mean SD Mean SD p

Comment ratinga 17 3.07 1.31 4.31 1.38 <0.001


Feeling thermometerb 1100 78.15 16.86 71.26 15.53 0.03
Semantic differentialb 19 5.84 1.05 5.25 1.07 <0.01
IAT Db n/a 0.27 0.41 0.41 0.43 0.05

IAT D = implicit association test; SD = standard deviation; n/a = not applicable.


a Higher ratings represent more prejudice toward Asians.
b Higher ratings represent greater affinity toward Asians.

3 SD above or below the mean; Field, 2009) were found in the comment ratings and
were therefore removed from the analyses.2
Five participants chose not to leave a comment on the Tertiary Education Website.
An independent samples t-test demonstrated that participants in the prejudiced social
norm condition (M = 4.31, SD = 1.38) expressed more prejudiced sentiments than
those in the antiprejudice condition; (M = 3.07, SD = 1.31), t(130) = 5.27, p < .001,
d = 0.92 (see Table 1). This supported the hypothesis that online comments placed by
other users directly impacts perceivers own comments.

Explicit prejudice
The two semantic differential items showed a strong internal consistency ( = .87)
and were therefore averaged into a composite score where larger numbers indicated
greater positivity toward Asians. The feeling thermometer was treated as a separate
index of prejudicial attitudes, where larger numbers also indicated less prejudicial
attitudes toward Asians, because the measures possessed different anchors.
The feeling thermometer revealed two extreme outliers and the semantic dif-
ferential ratings revealed one extreme outlier (i.e., 3 SD above or below the mean),
which were removed from their respective analysis (Field, 2009).3 A comparison of
the feeling thermometer scores revealed that participants exposed to prejudiced com-
ments (M = 71.26, SD = 15.53) showed less positive feelings toward Asians in general
than those exposed to antiprejudice online comments, (M = 78.15, SD = 16.86),
t(120) = 2.17, p = .03, d = 0.39. Similarly, semantic differential ratings were also
more negative among those exposed to prejudiced comments placed by other users
(M = 5.25, SD = 1.07) than those who were exposed to antiprejudiced comments
(M = 5.84, SD = 1.05), t(130) = 3.34, p = .001, d = 0.58, indicating that participants
in the prejudice norm condition showed less favorable feelings toward Asians relative
to those in the antiprejudice norm condition.
To further strengthen the findings, a MANOVA was conducted with both explicit
prejudice scales and the effects were still significant (p < .05), indicating that bias

566 Human Communication Research 41 (2015) 557576 2015 International Communication Association
M. Hsueh et al. Online Comments on Prejudice Expression

across semantic differential and feeling thermometers were each affected by preju-
diced and nonprejudiced social norms.

Implicit prejudice
An IAT D score was created using the algorithm described by Greenwald et al. (2003).
As detailed before, larger numbers on the IAT D score indicate stronger associations
between White/Good and Asian/Bad (i.e., greater negativity toward Asians and pos-
itivity toward Whites). An independent samples t-test revealed that participants in
the prejudiced norm condition (M = 0.41, SD = 0.43) demonstrated higher levels of
implicit prejudice toward Asians than participants in the antiprejudice norm condi-
tion, (M = 0.27, SD = 0.41), t(134) = 1.96, p = .05, d = 0.34. The IAT D scores indi-
cated that participants implicit prejudice toward Asians became significantly more
negative after being exposed to prejudiced comments relative to antiprejudiced com-
ments placed by other users.4

Discussion
This research examined the extent to which online comments placed by others estab-
lish a prejudiced or antiprejudiced social norm that in turn impacts perceivers per-
sonal attitudes and behaviors toward a social group. Results revealed that being in
an online environment where other readers have placed prejudiced or antiprejudiced
comments increased the extent to which participants own comments became biased
in the direction of their norm condition. Moreover, exposure to such prejudiced (or
antiprejudiced) online comments became generalized to impact not only perceivers
conscious attitudes toward an ethnic group, but also their unconscious or automatic
attitudes toward the entire group.

Contributions
This research makes several theoretical contributions to the literature from commu-
nication and social psychology. Although participants only read online comments
and did not interact with any of the fictitious commenters, participants still adopted
the groups social norm and adjusted their immediate response. The openended
approach to this study of social influences online allowed for freedom of expressions
by the participants more akin to real-world online settings and the use of qualita-
tive prejudice expression measures (in the form of online comments) provides an
extension of previous work in the area, which has mainly focused on external prej-
udice measures (Legault, Gutsell, & Inzlicht, 2011), making a novel contribution to
the literature. This finding was interesting as it suggested that the online comments
people come across can actually impact their own attitudes and behavior. Our findings
were consistent with the literature on persuasion in online contexts, which suggested
that consumers place heavy emphasis on others opinions when making judgments
(Ardelet & Briar, 2011; Cabezudo et al., 2013; Houston et al., 2011; Jansen et al., 2009;
Metzger et al., 2010).

Human Communication Research 41 (2015) 557576 2015 International Communication Association 567
Online Comments on Prejudice Expression M. Hsueh et al.

In addition to assessing qualitative (anti)prejudice expression measures (in the


form of online comments), the assessment of explicit attitudes toward an out-group
suggested that the online comments people read can form normative influence that
carry over beyond the immediate context. These results were similar to previous stud-
ies in which briefly hearing opinions of other in-group members were sufficient to
influence an individuals prejudicial attitudes (Blanchard et al., 1991, 1994).
To account for potential social desirability effects in explicit participant responses,
this study also included an implicit reaction-time measure that helps our under-
standing of the impact of online comments on implicit or unconscious biases toward
social groups. This research makes a unique theoretical contribution to the literature
by demonstrating that exposure to biased online comments influences readers to
adopt the online groups biases and those attitudes can linger subconsciously into
later explicit and implicit intergroup contexts.
Although this work demonstrated how unmoderated online comments may
increase the biases of online users, it also showed promise for online norms in prej-
udice reduction. Previous prejudice reduction interventions using external sources
have revealed mixed results by sometimes increasing discrimination instead of
reducing discrimination due to people rebelling against what some perceive to be a
violation of their autonomy (Legault et al., 2011), or require extensive practice and
time commitment (Kawakami, Dovidio, Moll, Hermsen, & Russin, 2000). The use of
social norms in this study allowed people to adopt differing prejudicial attitudes in a
subtle and autonomous manner.
These findings supported Blairs (2002) review of the literature on malleability of
implicit stereotypes and attitudes, which suggested that implicit biases are not fixed
and are susceptible to influences from the social environment. This research theoreti-
cally enriched this literature by showing that being exposed to prejudiced, as opposed
to antiprejudiced comments, placed by other users in an online setting can also impact
perceivers own feelings toward an ethnic group and this tendency was observed con-
sciously as well as unconsciously.

Limitations and future directions


Whereas the studys findings highlighted how something as trivial as the comments
placed by other online users may impact our own attitudes and expressions, there were
some limitations to consider that provide avenues for future work. This study lacked
a control group, making it difficult to determine whether participants only showed
increased prejudicial attitudes (e.g., Blanchard et al., 1994), or if the biases of partici-
pants became more polarized in both antiprejudiced and prejudiced directions (e.g.,
Blanchard et al., 1991). Although previous research on online comments suggests that
individual biases become polarized in both directions (Houston et al., 2011; Jansen
et al., 2009) whether that was the case here remains to be seen.
In addition, this study only investigated these effects within an hour long session
making it hard to know how long these effects last. Although it is conceivable that the
participants in the prejudice condition may have chosen to conform to the groups

568 Human Communication Research 41 (2015) 557576 2015 International Communication Association
M. Hsueh et al. Online Comments on Prejudice Expression

prejudiced or antiprejudiced comments because of normative influence, one would


expect that these participants prejudicial attitudes would revert to their original state
in the subsequent explicit and implicit measures; however, this was not the case.
This suggests that such online comments may have a small lingering effect on the
individual even after they have left the online context. Previous research has demon-
strated the malleability of implicit attitude change for at least 24 hours (Bicchieri &
Lev-On, 2007; Dasgupta & Greenwald, 2001). Therefore future studies can examine
the longevity of these effects and whether the lingering prejudicial attitudes can affect
subsequent intergroup attitudes and behavior even hours later. Our finding differs
from face-to-face group influences in that adoption of one groups norms does not
generalize into their interactions with another group (Postmes et al., 2000). Therefore
future studies can examine the longevity of these effects and whether the lingering
prejudicial attitudes can affect subsequent group interactions.
Another point to consider is that the predetermined comments in each of our
conditions were either antiprejudiced or prejudiced, creating an echo chamber-like
environment. Past research shows that consensus heuristics lead to increased power
of information and endorsement (Jansen et al., 2009; Metzger et al., 2010; Walther
& Jang, 2012), enhanced information credibility (Metzger et al., 2010; Thorson et al.,
2010) and as a validity cue for accuracy (Jansen et al., 2009; Petty, Priester, & Brinol,
2002). Woong-Yun and Park (2011) showed that being in the majority viewpoint
influenced an individuals willingness to comment as compared with being in the
minority viewpoint, thus creating a potential self-fulfilling climate for the majority
opinion.
The participants who chose not to comment (only a few in this study) may have
been unwilling to place dissenting comments for fear of ostracism among their peers.
Spears et al. (2002) found that individuals who had opposing viewpoints from the
group norm would be unsupported by other members of the group and would even-
tually fall in line with the majority group norm if left without support. Future studies
could therefore determine adoption of group identity by asking how identified the
participants feel toward their group. Although it was made clear in the information
form that participation was completely voluntary, only a few participants remained
silent or neutral. This may have been due to the online site explicitly asking par-
ticipants to provide feedback, although no more overt than blogs asking for viewer
feedback. However, without more data, it is hard to ascertain whether silence signifies
agreement or disagreement (DeAndrea, 2012). Future studies may wish to allow for
this by allowing participants to respond through nonpublic comments or by reducing
the expectation of participants to comment.
In real-world settings readers tend to choose like-minded media outlets, whether
it is specific Websites, blogs, or social network groups (Majchrzak et al., 2013). It is
worth considering that these personalized Websites can have a warranting effect
(Walther & Parks, 2002) where a small number of authors may audit the comments
that appear on their page so that only comments that the Website owner agrees with
are kept while dissenting comments are deleted (DeAndrea, 2012; Walther & Jang,

Human Communication Research 41 (2015) 557576 2015 International Communication Association 569
Online Comments on Prejudice Expression M. Hsueh et al.

2012) or completely blocked out (De Koster & Houtmann, 2008), thus creating an
echo chamber effect. Although more open-ended media Websites such as news sites
tend to be neutral, the comments placed by readers can be unpredictable, and in such
cases most readers would not know ahead of time what kind of comments to expect
from other users.
Nonetheless, we acknowledge that including only antiprejudiced or preju-
diced comments is a limitation of this study and future research should consider
implementing conditions with varying proportions of mixed comments (e.g., 80%
prejudiced comments, 20% antiprejudiced, or neutral comments vs. 60% prejudiced
comments vs. 40% antiprejudiced or neutral comments) to examine what level of
agreement within the online group is necessary for social norms to influence attitudes
and behavior.
Although comments are generally reactive in nature and commenters rarely
respond to replies to their comments (Walther et al., 2010), future studies could
also determine how individuals react to online comments. Moreover, studies could
investigate whether responses to participants antiprejudiced or prejudiced com-
ments would cause participants to strengthen their prejudicial stance, change to
adopt the attitudes of the responding commenter, or ignore the follow up completely.
Another avenue for future work would be to examine if the extent to which online
users carefully read and consider others prejudiced or antiprejudiced comments
influences attitude and behavior change in line with the social norms. Participants
who spend the most time reading others comments may although show the greatest
change in their personal biases.
While the findings of this study suggested that the empowerment provided by
the comfort and lack of accountability did not uniformly causes online commenters
to respond in antisocial ways, whether feelings of anonymity in online setting lead
participants to show more or less prejudice depending on the social norms remains
unclear. Future work may wish to examine whether a lack of accountability (dein-
dividuation) versus a sense of being identifiable (individuation) in an online setting
interacts with social norms (Reicher et al., 1995) to impact prejudicial attitudes and
behavior.
Finally, although the magnitude of the attitudinal differences between the two
norm conditions was statistically significant, the practical implications are unclear.
Previous research suggested that both implicit and explicit attitudes are valid pre-
dictors of behavior (Greenwald et al., 2009), and indeed, our participants comments
corresponded to the prevailing social norm. Future research can investigate the influ-
ence of attitude malleability on behavior, and the impact of spending an extended
period of time in online environments where prejudiced comments may be present.

Practical implications
This investigation suggested that online social norms could be looked upon as a
potential method for inducing prosocial attitudes and behavior. Hogg and Reid
(2006) asserted that an individuals identity can be molded by those around them. As

570 Human Communication Research 41 (2015) 557576 2015 International Communication Association
M. Hsueh et al. Online Comments on Prejudice Expression

the process of placing a comment requires participants to adhere to set procedures


(e.g., reading a comment, thinking of a response, typing that response, and then post-
ing the comment), it indicates that comments are not simply a slip of the tongue
(DeAndrea, 2012). The results of this study suggested that social norms could affect
an individual on a subconscious level, suggesting that an individuals identity could
be reshaped toward a prosocial identity through antiprejudiced online comments.
Another benefit of using this subtle normative approach to reducing prejudice is
the relatively low investment required. This study showed support for social norms
as an effective and efficient method for prejudice reduction, as multistep prejudice
reduction interventions are not always effective (Dixon, Durrheim, & Tredoux, 2007).
However, providers of online content also need to be aware of the potentially negative
influence online environments can facilitate. The inclusion of community managers
on sites such as the New York Times may be an efficient method for maintaining a
constructive environment, while still allowing readers to participate in dynamic ways.
An alternative to mitigate potentially negative influences between users may be to
offer a variety of user comments from which to select. Although users can still choose
unsupportive comments, these comments could be framed in less hateful and more
constructive ways, similar to restrictions placed on hate speech in various avenues of
everyday life.
Our findings suggested that by fostering a group norm where prejudice and stereo-
types about others are prevalent, members of that group were more likely to display
and carry over these negative sentiments in their everyday lives. On the other hand,
our findings also suggested that antiprejudiced norms can foster prosocial behavior.
The responsibility falls on users and proprietors to promote and maintain construc-
tive environments free from uncivil remarks. An optimistic outlook can be derived
from this study in terms of how malleable peoples attitudes are, and how relatively
simple it could be to promote antiprejudiced or prosocial attitudes.
However, a more pessimistic message is that online users can be negatively
impacted by the antisocial and harmful messages left by other users. As members of
the online community, we are all then responsible for promoting prosocial norms
in online environments. The challenge is, however, to make online users vigilant of
the potential impact their comments can have on others and how others comments
may impact them. Our research offers some insight into the indirect benefits and
downsides of Web 2.0. So the next time a Website asks you to leave your thoughts and
comments below, you may be wise to consider the surprising influence of something
as simple as your online comment.

Notes
1 Roughly equal numbers of participants used a name or a code (n = 64 vs. n = 72). We
tested if there were any differences in prejudice expression across these two groups and
found no such differences (all ps > .27).
2 Including the outliers in our analyses only further strengthens the effect.
3 Including the outliers in our analyses only strengthens the effect.

Human Communication Research 41 (2015) 557576 2015 International Communication Association 571
Online Comments on Prejudice Expression M. Hsueh et al.

4 Although the DVs were related in that they looked to examine prejudice toward Asians,
explicit and implicit measures are generally analyzed separately as they capture different
sentiments and are on very different indexes (see Gawronski & Bodenhausen, 2006;
Greenwald et al., 2009, 1998).

References
Anderson, A. A., Brossard, D., Scheufele, D. A., Xenos, M. A., & Ladwig, P. (2013). The
Nasty Effect: Online incivility and risk perceptions of emerging technologies. Journal of
Computer-Mediated Communication, 19, 373387. doi:10.1111/jcc4.12009.
Ardelet, C., & Briar, B. (2011). Influence on the recommendations of internet users: The role
of social presence and expertise. Recherche et Applications en Marketing, 26, 4567.
Bicchieri, C., & Lev-On, A. (2007). Computer-mediated communication and cooperation in
social dilemmas: An experimental analysis. Politics, Philosophy & Economics, 6, 139168.
doi:10.1177/1470594x07077267.
Blair, I. V. (2002). The malleability of automatic stereotypes and prejudice. Personality and
Social Psychology Review, 6, 242261. doi:10.1207/s15327957pspr0603_8.
Blanchard, F. A., Crandall, C. S., Brigham, J. C., & Vaughn, L. (1994). Condemning and
condoning racism: A social context approach to interracial settings. Journal of Applied
Psychology, 79, 993997. doi:10.1037/0021-9010.79.6.993.
Blanchard, F. A., Lilly, T., & Vaughn, L. A. (1991). Reducing the expression of racial prejudice.
Psychological Science, 2, 101105. doi:10.1111/j.1467- 9280.1991.tb.00108.x.
Brisbane, A. S. (2010, December 11). Readers with plenty to say. The New York Times.
Retrieved from http://www.nytimes.com/2010/12/12/opinion/12pubed.html.
Cabezudo, R. S. J., Izquierdo, C. C., & Pinto, J. R. (2013). The persuasion context and results
in online opinion seeking: Effects of message and source-The moderating role of network
managers. Cyberpsychology, Behavior and Social Networking, 16, 828835.
doi:10.1089/cyber.2011.0647.
Cialdini, R. B., & Goldstein, N. J. (2004). Social influence: Compliance and conformity.
Annual Review of Psychology, 55, 591621. doi:10.1146/annurev.55.090902.
Crandall, C. S., Eshleman, A., & OBrien, L. (2002). Social norms and the expression and
suppression of prejudice: The struggle for internalization. Journal of Personality and Social
Psychology, 82, 359378. doi:10.1037/0022-3514.82.3.359.
Dasgupta, N. (2004). Implicit ingroup favoritism, outgroup favoritism, and their behavioral
manifestation. Social Justice Research, 17, 143169. doi:10.1023/B:SORE.0000027407.
70241.15.
Dasgupta, N. (2009). Mechanisms underlying the malleability of implicit prejudice and
stereotypes: The role of automaticity versus cognitive control. In T. Nelson (Ed.),
Handbook of prejudice, stereotyping, and discrimination (pp. 267284). Mahwah, NJ:
Erlbaum.
Dasgupta, N., & Greenwald, A. G. (2001). On the malleability of automatic attitudes:
Combating automatic prejudice with images of admired and disliked individuals. Journal
of Personality and Social Psychology, 81, 800814.
De Koster, W., & Houtmann, D. (2008). Stormfront is like a second home to me: On virtual
community formation by right-wing extremists. Information, Communication & Society,
11, 11551176. doi:10.1080/13691180802266665.

572 Human Communication Research 41 (2015) 557576 2015 International Communication Association
M. Hsueh et al. Online Comments on Prejudice Expression

DeAndrea, D. C. (2012). Participatory social media and the evaluation of online behavior.
Human Communication Research, 38, 510528. doi:10.1111/j.1468-2958.2012.01435.x.
Deutsch, M., & Gerard, H. B. (1955). A study of normative and informational social influence
upon individual judgment. Journal of Abnormal and Social Psychology, 51, 629636.
doi:10.1037/h0046408.
Dixon, J., Durrheim, K., & Tredoux, C. (2007). Intergroup contact and attitudes toward the
principle and practice of racial equality. Psychological Science, 18(10), 867872.
Dylko, I., & McCluskey, M. (2012). Media effects in an era of rapid technological
transformation: A case of user-generated content and political participation.
Communication Theory, 22, 250278. doi:10.1111/j.1468-2885.2012.01409.x.
Edwards, C., Edwards, A., Qing, Q., & Wahl, S. T. (2007). The influence of computer-mediated
word-of-mouth communication on student perceptions of instructors and attitudes
toward learning course content. Communication Education, 56, 255277. doi:10.1080/
03634520701236866.
Fazio, R. H., & Olson, M. A. (2003). Implicit measures in social cognition research: Their
meaning and use. Annual Review of Psychology, 54, 297327. doi:10.1146/annurev.psych.
54.101601.145225.
Field, A. (2009). Discovering statistics using SPSS. London, England: Sage.
Gawronski, B., & Bodenhausen, G. V. (2006). Associative and propositional processes in
evaluation: An integrative review of implicit and explicit attitude change. Psychological
Bulletin, 123, 692731. doi:10.1037/0033-2909.132.5.692.
Glick, P., & Fiske, S. T. (1996). The ambivalent sexism inventory: Differentiating hostile and
benevolent sexism. Journal of Personality and Social Psychology, 70, 491512.
doi:10.1037/0022-3514.70.3.491.
Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual
differences in implicit cognition: The implicit association test. Journal of Personality and
Social Psychology, 74, 14641480. doi:10.1037/0022-3514.74.6.1464.
Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the implicit
association test: I. An improved scoring algorithm. Journal of Personality and Social
Psychology, 85, 197216. doi:10.1037/0022-3514.85.2.197.
Greenwald, A. G., Poehlmann, T. A., Uhlmann, E. L., & Banaji, M. R. (2009). Understanding
and using the implicit association test: III. Meta-analysis of predictive validity. Journal of
Personality and Social Psychology, 97, 1741. doi:10.1037/a0015575.
Hewstone, M., Islam, M. R., & Judd, C. M. (1993). Models of crossed categorization and
intergroup relations. Journal of Personality and Social Psychology, 64, 779793. doi:10.
1037/0022-3514.64.5.779.
Ho, C., & Jackson, J. W. (2001). Attitudes toward Asian Americans: Theory and measurement.
Journal of Applied Social Psychology, 31, 15531581. doi:10.1111/j.1559-1816.2001.
tb02742.x.
Hogg, M. A., & Reid, S. A. (2006). Social identity, self-categorization, and the communication
of group norms. Communication Theory, 16, 730. doi:10.1111/j.1468-2885.2006.00003.x.
Houston, J. B., Hansen, G. J., & Nisbett, G. S. (2011). Influence of user comments on
perceptions of media bias and third-person effect in online news. Electronic News, 5,
7992. doi:10.1177/1931243111407618.
Hughes, M. G., Griffith, J. A., Zeni, T. A., Arsenault, M. L., Cooper, O. D., Johnson, G., ...
Mumford, M. D. (2014). Discrediting in a message board forum: The effects of social

Human Communication Research 41 (2015) 557576 2015 International Communication Association 573
Online Comments on Prejudice Expression M. Hsueh et al.

support and attacks on expertise and trustworthiness. Journal of Computer-Mediated


Communication, 19, 325341. doi:10.1111/jcc4.12077.
Hughey, M. W., & Daniels, J. (2013). Racist comments at online news sites: A methodological
dilemma for discourse analysis. Media, Culture & Society, 35, 332347. doi:10.1177/
0163443712472089.
Internet Live Stats. (2014). Twitter usage statistics. Retrieved September 22, 2014, from http://
www.internetlivestats.com/twitter-statistics/.
Jansen, B. J., Zhang, M., Sobel, K., & Chowdury, A. (2009). Twitter power: Tweets as
electronic word of mouth. Journal of the American Society for Information Science and
Technology, 60, 21692188. doi:10.1002/asi.21149.
Jurkowitz, M (2014, March 26). The growth in digital reporting: What it means for
journalism and news consumers. Pew Research. Retrieved from http://www.
journalism.org/2014/03/26/the-growth-in-digital-reporting/.
Kawakami, K., Dovidio, J. F., Moll, J., Hermsen, S., & Russin, A. (2000). Just say no (to
stereotyping): Effects of training in the negation of stereotypic associations on stereotype
activation. Journal of Personality and Social Psychology, 78, 871888. doi:10.1037/0022-
3514.78.5.871.
Lapinski, M. K., & Rimal, R. N. (2005). An explication of social norms. Communication
Theory, 15, 127147. doi:10.1093/ct/15.2.127.
Lee, E.-J. (2006). When and how does depersonalization increase conformity to group norms
in computer-mediated communication? Communication Research, 33, 423447. doi:10.
1177/0093650206293248.
Lee, J. (2014, February 4). Coca-Cola Super Bowl ad: Can you believe this reaction? Retrieved
from http://www.usatoday.com/story/news/nation-
now/2014/02/03/coca-cola-ad-super-bowl-racism/5177463/.
Legault, L., Gutsell, J. N., & Inzlicht, M. (2011). Ironic effects of antiprejudice messages: How
motivational interventions can reduce (but also increase) prejudice. Psychological Science,
22, 14721477. doi:10.1177/0956797611427918.
Majchrzak, A., Faraj, S., Kane, G. C., & Azad, B. (2013). The contradictory influence of social
media affordances on online communal knowledge sharing. Journal of Computer-
Mediated Communication, 19, 3855. doi:10.1111/jcc4.12030.
McLaughlin, C., & Vitak, J. (2011). Norm evolution and violation on Facebook. New Media &
Society, 14, 299315. doi:10.1177/1461444811412712.
Metzger, M. J., Flanagin, A. J., & Medders, R. B. (2010). Social and heuristic approaches to
credibility evaluation online. Journal of Communication, 60, 413439. doi:10.1111/
j.1460-2466.2010.01488.x.
Nosek, B. A., Smyth, F. L., Hansen, J. J., Devos, T., Lindner, N. M., Ranganath, K. A., ... Banaji,
M. (2007). Pervasiveness and correlates of implicit attitudes and stereotypes. European
Review of Social Psychology, 18, 3688. doi:10.1080/10463280701489053.
Petty, R. E., Priester, J. R., & Brinol, P. (2002). Mass media attitude change: Implications of the
elaboration likelihood model of persuasion. In J. Bryant & D. Zillmann (Eds.), Media
effects: Advances in theory and research (pp. 155198). Mahwah, NJ: Erlbaum.
Plant, E. A., & Devine, P. G. (1998). Internal and external motivation to respond without
prejudice. Journal of Personality and Social Psychology, 75, 811832. doi:10.1037/0022-
3514.75.3.811.

574 Human Communication Research 41 (2015) 557576 2015 International Communication Association
M. Hsueh et al. Online Comments on Prejudice Expression

Popular Science. (2013). Why were shutting off our comments. Retrieved September 24,
2013, from http://www.popsci.com/science/article/2013-09/why-were-shutting-
our-comments?
Postmes, T., & Spears, R. (2002). Behavior online: Does anonymous computer
communication reduce gender inequality? Personality and Social Psychology Bulletin, 28,
10731083. doi:10.1177/01461672022811006.
Postmes, T., Spears, R., & Lea, M. (2000). The formation of group norms in computer-
mediated communication. Human Communication Research, 26, 341371. doi:10.1093/
hcr/26.3.341.
Reicher, S. D., Spears, R., & Postmes, T. (1995). A social identity model of deindividuation
phenomena. European Review of Social Psychology, 6, 161198. doi:10.1080/14792779
443000049.
Rimal, R. N., & Real, K. (2003). Understanding the influence of perceived norms on
behaviors. Communication Theory, 13(2), 184203.
Sagar, H. A., & Schofield, J. W. (1980). Racial and behavioral cues in black and white
childrens perceptions of ambiguously aggressive acts. Journal of Personality and Social
Psychology, 39, 590598. doi:10.1037/0022-3514.39.4.590.
Schultz, P. W., Nolan, J. M., Cialdini, R. B., Goldstein, N. J., & Griskevicius, V. (2007). The
constructive, destructive, and reconstructive power of social norms. Psychological Science,
18, 429434. doi:10.1111/j.1467-9280.2007.01917.x.
Sinclair, S., Lowery, B. S., Hardin, C. D., & Colangelo, A. (2005). Social tuning of automatic
racial attitudes: The role of affiliative motivation. Journal of Personality and Social
Psychology, 89(4), 583.
Spears, R., Lea, M., Corneliussen, R. A., Postmes, T., & Ter Harr, W. (2002).
Computer-mediated communication as a chance for social resistance: The strategic side
of SIDE. Small Group Research, 33, 555574. doi:10.1177/104649602237170.
Stangor, C., Sechrist, G., & Jost, J. T. (2001). Changing racial beliefs by providing consensus
information. Personality and Social Psychology Bulletin, 27, 486496. doi:10.1177/
0146167201274009.
Stuff. (2013, May 12). Chinese cheats rort NZ Universities with fakes. Retrieved June 1, 2013
from http://www.stuff.co.nz/national/education/8662224/Chinese-cheats-rort-
NZ-universities-with-fakes 12/05/2013.
Thorson, K., Vraga, E., & Ekdale, B. (2010). Credibility in context: How uncivil online
commentary affects news credibility. Mass Communication and Society, 13, 289313.
doi:10.1080/15205430903225571.
Toosi, N. R., Babbitt, L. G., Ambady, N., & Sommers, S. R. (2012). Dyadic interracial
interactions: A meta-analysis. Psychological Bulletin, 138, 127. doi:10.1037/a0025767.
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, 469492. doi:10.1111/j.1468-2958.2010.01384.x.
Walther, J. B., & Jang, J.-W. (2012). Communication processes in participatory websites.
Journal of Computer-Mediated Communication, 18, 215. doi:10.1111/j.1083-6101.
2012.01592.x.
Walther, J. B., & Parks, M. R. (2002). ). Cues filtered out, cues filtered in: Computer-mediated
communication and relationships. In M. L. Knapp & J. A. Daly (Eds.), Handbook of
interpersonal communication (pp. 529563). Thousand Oaks, CA: Sage.

Human Communication Research 41 (2015) 557576 2015 International Communication Association 575
Online Comments on Prejudice Expression M. Hsueh et al.

Woong-Yun, G., & Park, S. Y. (2011). Selective posting: Willingness to post a message online.
Journal of Computer-Mediated Communication, 16, 201227. doi:10.1111/j.1083-6101.
2010.01533.x.
Yanovitzky, I., & Rimal, R. N. (2006). Communication and normative influence: An
introduction to the special issue. Communication Theory, 16, 16. doi:10.1111/j.
1468-2885.2006.00002.x.
Zitek, E. M., & Hebl, M. R. (2007). The role of social norm clarity in the influenced
expression of prejudice over time. Journal of Experimental Social Psychology, 43,
867876. doi:10.1016/j.jesp.2006.10.010.

576 Human Communication Research 41 (2015) 557576 2015 International Communication Association

S-ar putea să vă placă și