Available online at www.sciencedirect.

com

ScienceDirect
Journal of Interactive Marketing 41 (2018) 44 – 59
www.elsevier.com/locate/intmar

Anxiety and Ephemeral Social Media Use in Negative

eWOM Creation
Lane T. Wakefield a,⁎& Robin L. Wakefield b
a
Mercer University, United States
b
Baylor University, United States

Abstract

Consumers regularly share negative eWOM regarding brand experiences, yet the effects of emotions in this process are unknown. Using the
framework of Stimulus–Organism–Response theory we operationalized our study in the context of ephemeral (self-deleting) social media. Across
two studies, consumers experienced anxiety during the creation of eWOM, particularly after considering a negative brand experience. Anxiety was
manifest in consumers' decisions to reduce the time availability of online messages and appeased by the use of ephemeral social media. This
behavior, in theory, reduces the threat to consumers' impression management goals caused by sharing negative eWOM within their social
networks. Further evidence is found as consumers limit (lengthen) message availability a) when including less (more) positive words in their
messages, or b) when sharing about non-economic (economic) experiences. We discuss the implications of consumer emotion and anxiety on the
availability of eWOM in social media.
© 2017 Direct Marketing Educational Foundation, Inc., dba Marketing EDGE.

Keywords: Anxiety; eWOM; Ephemeral social media; Emotion; Negativity; Impression management; Snapchat; Instagram

Introduction in our investigation of the role of consumer anxiety in online

Consumers often share emotional experiences in social media,
including those stimulated by brands. One's favorite sports team
losing may lead to sadness, an electronic device with a faulty part
may lead to frustration, or an unpleasant interaction with a
customer service representative may promote anger and the desire
to share the experience. When consumers encounter negative brand
experiences and subsequent emotions, they respond with coping
behavior (Duhachek 2005). Coping behavior may include creating
electronic word-of-mouth (eWOM) to complain (Stephens and
Gwinner 1998) or voicing sentiment to gain benefits such as social
support (Hennig-Thurau et al. 2004). While emotions that motivate
and follow eWOM have been studied, the observation of emotion
in the act of creating eWOM is a missing element. We address this
⁎ Corresponding author.
E-mail addresses: Wakefield_lt@mercer.edu (L.T. Wakefield),
Robin_wakefield@baylor.edu (R.L. Wakefield).
sharing about negative brand experiences.
Anxiety, or feelings of anxiousness, apprehension or worry
(Brooks and Schweitzer 2011) are widespread in online environ-
ments (Caplan 2007; Correa, Hinsley, and De Zuniga 2010),
particularly among social media users (Andreassen et al. 2016).
Anxiety is a negative emotional response to a stressor or threat
(Lazarus and Folkman 1984). In line with the literature, we define
social media as an interactive communication platform for content
creation and sharing within social networks (VanMeter, Grisaffe,
and Chonko 2015). The interactive component suggests anxiety is
salient when creating eWOM because individuals who share
content online expect their posts to be scrutinized by their
audiences (Krämer and Winter 2008), thus the prospect of
scrutiny produces anxiety as individuals compose messages.
Furthermore, we suggest the creation and sharing of negative
eWOM (i.e., eWOM containing negative sentiment) is particularly
anxiety producing because negative messages are likely to conflict
with individuals' online impression management goals. The
need for self-presentation as part of impression management

https://doi.org/10.1016/j.intmar.2017.09.005
1094-9968© 2017 Direct Marketing Educational Foundation, Inc., dba Marketing EDGE.
 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59
(Berger 2014) is typical for consumers (Fiske 2001) and may be
amplified online (Suler 2004). Thus, anxiety would be prevalent
when sharing negative eWOM especially as negative information
generally weighs more heavily on judgements (Anderson 1965;
Hancock and Dunham 2001).
We operationalize our study in the context of ephemeral social
media for two purposes. Ephemeral social media are online
communication platforms in which the user predetermines the
lifespan of a message/post, or message availability. One objective
is to explore how anxiety and negativity manifest in consumers'
choice of message availability. The other goal is to provide
insights into consumer behavior for this emerging genre of social
media that to date has received little inclusion in empirical
studies. In an ephemeral social media feed, a user creates eWOM
that will be erased automatically or after a designated time
predetermined by the user. Snapchat is the most valuable of those
dedicated to ephemeral messaging with a recent valuation of $20
billion (Hartung 2016). Facebook and its subsidiary, Instagram,
also offer ephemeral features (Wagner 2017). Snow, a startup
ephemeral messaging system in Asia, claimed up to 50 million
active users in its first year (Russell 2017). Some mobile
applications, such as Xpire, work in tandem with traditional
social media (e.g., Facebook) to allow the user to set a
self-destruct timer on content.
We investigate two research questions regarding consumers'
expression of eWOM following a brand experience. These
include: 1) How does a negative brand experience influence
consumers' anxiety while they are in the act of creating
eWOM? 2) How do consumers cope with negativity and
anxiety when using ephemeral social media? To address these
questions, we created a research model using Stimulus–
Organism–Response (S–O–R) theory as a framework. Across
two studies involving several product categories and brand
experience types, we examine the effect of a brand experience
(stimulus) on individuals' felt anxiety and expression of
negative sentiment (organism). We then examine the subse-
quent effects of anxiety and negativity on message availability
preferences and individuals' post-anxiety (response). The
results indicate that consumers experience anxiety as they
compose eWOM in response to a negative brand experience.
This anxiety, in turn, tends to influence message availability
when using ephemeral social media. We also find evidence that
online impression management goals may interfere with the full
expression of negative sentiment in eWOM because consumers
may include more positive words to alleviate the conflict, or
else shorten message availability. Further, consumers' message
availability preferences differ according to the type of brand
experience (economic, non-economic). Overall, ephemeral
systems tend to relieve consumer anxiety associated with
negative eWOM benefitting both consumers and brands.
While researchers have investigated negative emotion resulting
from incidents and antecedents to complaining (Chebat, Davidow,
and Codjovi 2005; Stephens and Gwinner 1998), an examination
of negative emotion throughout a given action and coping process
is lacking. We observe the state of negative emotion in-process
and after action is taken to understand users' coping behavior
related to its impact on eWOM. Prior research indicates that as
45
many as 71% of Facebook users may engage in self-censorship
during the creation of new posts or comments to posts (Madrigal
2013). Our study provides insight into why and how eWOM is
constrained. Given the strong evidence of emotion (e.g., anxiety)
in social media use and the considerable number of individuals
(about 50%) whose purchase decisions are influenced by online
opinion (Mathwick and Mosteller 2017), user emotion during
content creation has implications for the usefulness of eWOM and
the value of social media platforms if content is restricted.
Additionally, as ephemeral social media becomes mainstream
(viz. hundreds of millions of users across platforms worldwide), it
is important for marketers to understand how consumers utilize its
primary feature (i.e., self-deleting messages) to carry out a
common task (i.e., creating and sharing eWOM). Little empirical
work has been conducted in ephemeral social media contexts due
to the difficulty of examining transient content (Piwek and Joinson
2016). Our study fills multiple gaps by exploring consumers'
negative emotion following a brand experience, the role of anxiety
in creating and sharing eWOM, and ephemeral technology use
behavior.
Conceptual Framework and Hypotheses
We operationalize negative eWOM as consumers' social
media messages related to a brand experience that contain
negative sentiment. Our focus on negative eWOM is not
limited to complaining. A large proportion of consumers
contact brands' customer service using social media (80%)
and a portion of those may complain (20%) (Morrison 2016).
However, complaining may occur for different purposes and
may be directed at multiple parties (viz. friends, strangers and
brands) even when the consumer is satisfied with the brand
(McGraw, Warren, and Kan 2014). Thus, we emphasize
negative eWOM rather than solely complaining behavior
because eWOM is a general indicator of customer engagement
(Van Doorn et al. 2010) that brands actively monitor and
encompasses more than complaints.
Mehrabian and Russell's (1974) S–O–R theory proposes that
the environment encompasses stimuli (S) that alter an
individual's internal disposition, or organism (O), that results in
a conscious outward response (R). Consumer behavior studies
have utilized S–O–R theory to investigate the influence
of psychological variables related to the environment on
individuals. For example, in a retail store environment, stimuli
(social, ambient and design characteristics) may alter consumers'
organismic state (level of arousal), leading to a response (making
a purchase) (Chang, Eckman, and Yan 2011). Other studies have
utilized S–O–R theory to investigate how individuals browse
websites (Eroglu, Machleit, and Davis 2001; Kim and Lennon
2013; Richard and Chandra 2005). S–O–R theory is applicable
to our study because one's experience with a brand is a relevant
stimulus to induce emotion regarding the brand (i.e., organismic
reaction) which may be expressed in eWOM (i.e., behavioral
response).
We created the research model in Fig. 1 to examine the effects
of brand experience on task-related emotion (i.e., anxiety when
creating eWOM), message negativity and the availability of a
46 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59
Fig. 1. Research model.
social media message about the experience. We also compare
consumer anxiety during the eWOM creation task with anxiety
following the completion of the task.
Brand Experience
A brand is defined as “a name, term, sign, symbol, or design,
or combination that is intended to identify the goods and services
of one seller or group of sellers and to differentiate them from
those of competitors” (Kotler 1997, p. 443). However, a brand
experience is broader and encompasses other factors such as
“specific sensations, feelings, cognitions, and behavioral re-
sponses triggered by specific brand-related stimuli” (Brakus,
Schmitt, and Zarantonello 2009, p. 53). In accord with S–O–R
theory, a brand-related object may act as a stimulus that induces
emotion, and emotion, in turn, may motivate a behavioral
response toward the brand. For example, individuals encounter
brands continuously from the moment their branded alarm sounds
to the drinking of branded coffee before putting on branded
clothing. Some brands may stay “invisible” without daily
recognition (e.g., coffee) (Coupland 2005), yet others may be
more overt (Batra, Ahuvia, and Bagozzi 2012) such as clothing
and cars. However, if the branded object fails expectations (e.g.,
alarm doesn't sound, car breaks down, coffee is bitter) then
negative attributions are likely directed toward the brand, rather
than to alarm clocks, cars or coffee in general.
A brand experience may encompass different types of
brand-related stimuli such as product features, services, and
performance associated with the brand. Product features are
attributes noticed by the consumer such as the soft texture or light
weight of specific athletic apparel — haptic features that can
influence product judgment (Peck and Childers 2003). These
stimuli may be accompanied by a service delivered by a cashier or
salesperson, which may also influence brand experience via the
customer's service experience (Berry 2005). From a broader
perspective, a company's overall brand performance may act as a
stimulus. Brand performance is defined as “the relative measure-
ment of the brand's success in the marketplace,” (O'Cass and Ngo
2007, p. 871), such as stock prices for public companies. Brand
performance may be evaluated in other ways such as the wins and
losses of professional sports teams. In sum, a variety of
brand-related stimuli affects individuals, each with the potential
to influence the internal disposition (i.e., organismic response) of
the individual toward the brand.
Social Media Message Negativity, Impression Management
and Anxiety
An individual's internal disposition is partially composed of
organismic responses to stimuli in an environment (Mehrabian and
Russell 1974); however, the intensity of the organismic response is
situationally dependent (Chandler and Lusch 2015). For example,
individuals' feelings about guns change dramatically after terrorist
attacks or when new gun laws are proposed (Newport 2016). More
directly, a poor product or service experience with a brand may
result in strong negative emotion after appraisal (Stephens and
Gwinner 1998). Thus, when an individual is stimulated by a
brand-related experience of sufficient intensity, a change in
emotion is more likely. We use a negative brand experience to
stimulate individuals' organismic response.
Individuals who have had a poor brand experience are likely
to share it online. Research indicates online negative commu-
nication is more common than face-to-face negative commu-
nication (Suler 2004) partly because online communication
may be perceived as less risky than offline (Lee and Stapinski
2012). Electronic messages are also more likely to reveal
personal thoughts (Adler and Adler 2008) and social media
users may consider their digital content an open diary (Marwick
 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59
and Boyd 2011). Given the widespread use of social media in
general (~ 78% of US consumers) (Statista 2015) and for
sharing information with and about brands (Morrison 2016), a
significant relationship between a negative brand experience
and social media message negativity is expected and included
in Fig. 1.
Additionally, composing eWOM about a poor brand experi-
ence is likely to contribute to social media task anxiety. Anxiety
includes feelings of anxiousness, apprehension or nervousness
(Brooks and Schweitzer 2011). Anxiety is prevalent in online
contexts (Caplan 2007; Correa, Hinsley, and De Zuniga 2010) and
may persist offline (Kross et al. 2013). In the psychology literature,
negative emotions such as anxiety arise when individuals' goals
are threatened (Lazarus 1991). Thus, we argue that the desire to
share about a negative brand experience online will conflict with
individuals' online impression management goals. Impression
management is illustrated by social media users who continually
render a positive portrait of themselves online (Wilcox and
Stephen 2012). Furthermore, self-enhancement (a facilitator of
impression management; Berger 2014) behavior is prevalent in
consumers (Fiske 2001), especially online (Suler 2004). When the
audience is greater than one, sharing positive messages typifies
self-presentation communication (Barasch and Berger 2014). In
sum, managing one's online impression (e.g., via positive
messaging) may conflict with the need to express negative
sentiment about an experience, creating anxiety.
Benefits are associated with both impression management and
eWOM activities which may contribute to users' conflict and
anxiety. For example, impression managers have more friends
(Krämer and Winter 2008), find more social support (Newman et
al. 2011), attract romantic partners (Ellison, Heino, and Gibbs
2006), and may be evaluated more highly by supervisors (Singh,
Kumra, and Vinnicombe 2002). These individuals create self-
presentation opportunities online (Krämer and Winter 2008) and
are more strategic online in presenting the ideal self (Ellison,
Heino, and Gibbs 2006). Thus, positivity in eWOM would be
important for impression managers. Yet, expressing negative
sentiment or complaining about a brand experience also has
benefits. Warning or entertaining others (McGraw, Warren, and
Kan 2014), venting (Sparks and Browning 2010; Ward and
Ostrom 2006), seeking redress (Blodgett, Wakefield, and Barnes
1995), damaging the brand, boosting self-worth, and gaining social
support (Hennig-Thurau et al. 2004) are typical rewards sought by
those who suffer a negative brand experience.
Some consumers may avoid the consequences of negative
expression by reframing comments to seem more positive
(Barasch and Berger 2014), using humor (McGraw, Warren,
and Kan 2014) or staying silent (Chebat, Davidow, and Codjovi
2005). However, many individuals facing an emotional
experience have a need to share it (Pennebaker, Zech, and
Rimé 2001), even under social constraints (Pennebaker 1993).
In social media, sharing behavior is more likely when messages
contain emotional content (Stieglitz and Dang-Xuan 2013). In
summation, the need to express negative sentiment amplifies
the conflict between doing so and maintaining a positive online
impression, resulting in greater anxiety for the task of writing
eWOM.
47
H1. Negative brand experiences will result in greater anxiety
when creating eWOM compared to positive brand experiences.
Coping with Negativity and Anxiety by Restricting Message
Availability
Responding to an emotional experience, or coping, is
necessary when negative emotions are present (Schwartz 1990).
Coping is defined as “the set of cognitive and behavioral
processes initiated by consumers in response to emotionally
arousing, stress inducing interactions with the environment aimed
at bringing forth more desirable emotional states and reduced
levels of stress” (Duhachek 2005, p. 42). Coping involves an
appraisal process (Stephens and Gwinner 1998) in which the
individual evaluates the significance of the experience relative to
his/her well-being. We suggest that coping with a poor brand
experience by creating negative eWOM in social media is likely
to result in an emotional trade-off. Emotional trade-offs occur
when gains from one action result in gains given up by foregoing
another action (Luce, Bettman, Payne 2001). On the one hand,
the need to share negative sentiment may be satisfied, while on
the other hand one's positive online impression is at risk. If
the social media audience contains a mixture of personal and
professional circles, the costs (i.e., risks) may increase (Marwick
and Boyd 2011). Additionally, beliefs about the relative
permanence of social media messages may exacerbate potential
risks.
The Internet has shifted interpersonal communication from,
“ephemeral, transient, unmappable, and invisible” to “permanent,
mappable, and viewable” (Manovich 2009, p. 324) because
computer-mediated communications are largely archived. The
Library of Congress has been cataloging even trivial Twitter
“tweets” since 2010 (Eversley 2011), leading to a New York
Times article by Rosen (2010), “The Web Means the End of
Forgetting.” Brad Paisley recently released a song titled
“Selfie#theinternetisforever” (Burchard 2017). However, inno-
vation has also led to the creation of ephemeral, or self-deleting,
online communication platforms. While early adopters purposed
ephemeral social media for sharing explicit content (Hasinoff
2013; Poltash 2012), that is no longer the norm (Roesner, Gill,
and Kohno 2014). Users of the most prominent ephemeral social
media platform, Snapchat, primarily share platonic everyday
experiences (Bayer et al. 2015).
Ephemeral social media platforms may be popular partly
because ephemerality facilitates coping with negative emotion
and the conflict related to managing impressions. For example,
limiting message availability could offer the perception of a
smaller audience size, which may be important for those
managing multiple audiences (Ellison et al. 2011) or with
concern for the content's future exposure. Hence, social media
users sharing negative eWOM are likely to desire more
stringent time restrictions on their messages. Additionally,
social media users' audiences generally come from multiple
social contexts (Marwick and Boyd 2011) which are likely to
have less interpersonal closeness (Dubois, Bonezzi, and De
Angelis 2016). Thus, we expect greater anxiety related to
composing negative eWOM will also lead to shorter message
48 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59

availability. In sum, the coping response related to negative of anger, sadness, or anxiety. The LIWC is valid for measuring
eWOM and task anxiety would be to limit the availability of expressed emotion (Kahn et al. 2007) and has been used in a
messages. variety of literature including marketing (Barasch and Berger
2014; Tausczik and Pennebaker 2010) and sport management
H2. The greater the negativity in a message, the shorter the
(Mankad, Gordon, and Wallman 2009). The LIWC also served
message availability.
as a check for the brand experience manipulation by observing
H3. The greater the task anxiety, the shorter the message the frequency of negative emotion words used in each message.
availability. Additional checks were included in the survey to ensure the
respondents understood the concept of ephemeral messaging
Study 1: Sample and Method and paid attention. Control variables accounted for variance in
message availability due to age, gender, income and social
A nationwide panel of U.S. participants (n = 164) was media network size in terms of the number of Twitter followers
recruited through Amazon's Mechanical Turk (MT) service. and Facebook friends. We also controlled for team identifica-
U.S. MT samples provide demographics equivalent to those of tion (Fisher and Wakefield 1998) due to the sports team
consumer panels and student samples, in contrast to non-U.S. context. Table 1 shows the descriptive statistics according to
MT samples (Steelman, Hammer, and Limayem 2014). Of the manipulation group (negative or control).
those disclosing gender, 62% were male with a mean age of Before proceeding with the structural equations modeling,
35 years. Annual incomes were generally less than $50,000 per we compared the data from the group composing a Twitter
year (65%), with 25% earning between $50,000 and $100,000 message with the data from the group composing a Facebook
and 6% earning more than $100,000. message. For the key constructs (task anxiety, message
In Study 1, the respondents indicated a favorite professional negativity, message availability), the means between the two
sports team from any of the professional leagues (National groups were not significantly different (task anxiety t = 0.797,
Hockey League, National Football League, National Basketball p = 0.426; message negativity t = − 1.672, p = 0.097; message
Association, Women's National Basketball Association, or availability t = − 0.444, p = 0.658) and the data were pooled to
Major League Baseball) and the selection was inserted into the test the research model.
survey scenario and survey questions as appropriate. Associating
consumers' current experience with a brand, such as their favorite Results
professional sports team, creates a realistic scenario that reduces
psychological distance (Trope, Liberman, and Wakslak 2007). Measurement Model
Two experience scenarios were repurposed from Wakefield and The model was tested using partial least squares structural
Bennett (in press) for this study. Each scenario was pretested to equation modeling (PLS-SEM) with SmartPLS 3.0 software
ensure the experience was significantly better or worse, (Ringle, Wende, and Becker 2015). PLS-SEM is the ideal choice
respectively. Negative Brand Scenario — “Imagine experiencing for exploratory studies, the analysis of latent variables, and for
the following event: You are at a [favorite team's] game and they research models that contain formative constructs (Hair et al.
are losing horribly.” Positive Control Scenario — “Imagine 2017). The appropriate sample size is important to achieve
experiencing the following event: You went with your friend to a sufficient power to detect significant relationships in PLS. For the
[favorite team's] game and it was really great.” After reading the research model including controls, achieving statistical power of
scenario, each subject was randomly assigned to write a message 80% and minimum R2 of .10 in the dependent variable at a 5%
about the experience for Twitter or Facebook. The subjects also significance level would require a sample size of 144 (Cohen
read a short passage describing an ephemeral messaging 1992). Our sample of 164 responses indicates sufficient power to
application and were instructed to set a timer on their social
media messages. Message availability was measured with a
7-point scale ranging from (1) minutes (1–60) to (7) I don't want
to delete this message/post (Wakefield and Bennett in press). Table 1
Descriptive statistics by manipulation for Study 1.
Participants indicated their felt anxiety related to composing the
message on a 4-item reflective scale (i.e., anxious, apprehensive, Variable Negative condition Control condition
worried and nervous; Brooks and Schweitzer 2011) ranging from N = 86 N = 78
(1) Not at all to (7) Completely. Mean SD Mean SD
We used the Linguistic Inquiry and Word Count (LIWC) Message negativity 2.15 2.85 0.201 .589
(Pennebaker, Francis, and Booth 2001) text analysis program to Task anxiety 2.22 1.38 1.72 1.29
measure the number of anger, sadness and anxiety words in Message availability 3.49 2.18 4.32 2.01
# Twitter followers 66.69 111.0 94.59 135.3
each respondent's social media message. These three variables
# Facebook friends 240.5 253.1 215.9 241.8
are formative items that comprise the message negativity Age 34.57 8.07 35.87 9.87
construct. Richins' (1997) identified eight negative emotions Income 4.79 3.69 4.46 2.97
(i.e., anger, discontent, worry, sadness, fear, shame, envy, and Gender 1.34 0.476 1.27 0.447
loneliness) and except for discontent, all are included in the Team identity 5.22 1.88 5.54 1.67
Ephemeral use 2.17 0.513 2.24 0.514
LIWC dictionary within descriptors for the negative emotions
 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59
detect small effects on the dependent variable, message
availability.
We used a 5,000 bootstrap resampling method from the
original data set as recommended (Hair et al. 2017). The reliability
and consistency of the reflective construct (task anxiety) were
evaluated with Cronbach's alpha and composite reliability (CR).
Both tests exceeded the 0.7 benchmark (Nunnally 1978) as shown
in Table 2. Likewise, the average variance extracted (AVE) for
convergent validity was above the 0.5 threshold. Each item for
task anxiety loaded highest on its associated factor with all
loadings above the minimum of 0.7 (Fornell and Larcker 1981).
Additionally, the heterotrait–monotrait (HTMT) ratios of item
correlations across all construct pairs ranged from .03 to .31, well
below the conservative .85 threshold for discriminant validity
(Hair et al. 2017). Descriptive statistics and correlations are in
Table 3.
We used the variance inflation factor (VIF) to evaluate the
formative items of the message negativity construct because
high correlations are generally not acceptable for formative
items. All VIF values were less than 2.0, lower than the 5.0
benchmark for excessive collinearity (Hair et al. 2010).
Additionally, the weight of each formative item was significant
at p b .001, with t-values ranging from 4.10 to 6.79. Anger had
the largest relative contribution to the message negativity
construct (weight = .715) followed by sadness (.580) and
anxiety (.424).
Structural Model
The structural model results and estimated path weights are
shown in Fig. 2. As expected, a negative brand experience is
significantly related to greater message negativity. H1 is supported
in that a negative brand experience (negative = 0, control = 1) is
significantly related to greater task anxiety (−.183, p b .01).
Message negativity is significantly related to message availability
(−.152, p b .05) supporting H2, and task anxiety is related to
message availability (−.175, p b .05) in support of H3. Respon-
dents using more negativity and those with more anxiety chose to
decrease the availability of social media messages. Task anxiety
was also significantly greater (t = 2.38, p = .01) for those in the
negative brand experience scenario (M = 2.22, SD = 1.37)
compared to the control condition (M = 1.72, SD = 1.29).
Table 2
Measurement model results.
Factor CR Cronbach's AVE
loading alpha
Task anxiety .94 .92 .81
How anxious did you feel while .895
completing the task?
How apprehensive did you feel .866
while completing the task?
How worried did you feel while .911
completing the task?
How nervous did you feel while .921
completing the task?
CR = composite reliability.
AVE = average variance extracted.
49
As a manipulation check, the frequency of negative emotion
words from those in the negative brand experience group (M =
5.52) was greater than those in the control group (M = 0.85).
This indicates the respondents in the negative experience
scenario were considering a negative brand experience when
writing social media messages. Examples of respondents'
messages with the most negative emotion words include the
following:
a. ‘This is the worst game ever. We're losing badly, I can't
believe they picked now to lose when I′m at the game’
b. ‘Arghhh! Dbacks down 8-0 bottom of the 9th, I blame
Obama!’
c. ‘My team is losing. I can't believe the offense is turning over
so many plays. There is no defense at all. What is wrong
with them. Really???’
d. ‘sitting at the #lakers game they are losing horribly,
#ByronScott needs to make changes’
e. ‘This sucks!!! I'm coming close to finding a new team. I hate
losers.’
The control variables indicate that gender and age
are not significantly related to message availability.
However, respondents' income (−.125, p b .05), Twitter
network (# followers) (.124, p b .05) and number of Facebook
friends (.185, p b .05) are significantly associated with
message availability. Team identification was not significantly
related to message availability (.030, p = .394). The effect size
(f 2 ) of message negativity on message availability is .026 and
task anxiety on message availability is .032, both indicating
small effects (Cohen 1992). The model explains about 11.2%
of the variance in message availability.
The blindfolding technique was used to calculate Stone–
Geisser's Q2 (Stone 1974) which is a criterion for the predictive
relevance of the PLS path model. The blindfolding technique and
an omission distance of seven were used to show the Q2 values
for the independent variables were greater than zero (message
negativity = .056; task anxiety = .023). This method indicates
that the research model accurately predicted data not used in the
model estimation; the research model has predictive relevance for
message availability.
Additional statistical analyses of three direct effect models
are provided in Table 4. In Model 1, little variance (R2 = 1.7%)
in message availability is explained by the controls: age,
gender, income and team identity. The addition of Twitter and
Facebook network size in Model 2 shows that message
availability is influenced by the number of Facebook friends
and explained variance increases to R2 = 5.9%. However, the
organismic response variables (message negativity and task
anxiety) in Model 3 account for significant additional variance
in message availability (R2 = 13.8%). The effect size (f2) of the
change in R2 between models 2 and 3 indicates a small to
medium effect (f2 = .09), using .02 and .15 as baselines for
small and medium effects, respectively (Cohen 1988). Overall,
the statistical results support 1) message negativity and task
anxiety as indicators of an organismic response to brand
50 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59

Table 3
Descriptive statistics and correlation table for Study 1.
Construct M SD 1 2 3 4 5 6 7 8 9
1. Message negativity 1.2 2.3 1.00
2. Task anxiety 1.98 1.36 .024 1.00
3. Message avail 3.88 2.13 −.105 −.136 1.00
4. # Twitter followers 79.9 123 −.035 .066 .146 1.00
5. # FB friends 228.8 247 .004 .023 .196 .310 1.00
6. Age 38.2 8.9 −.023 .047 −.008 −.103 −.153 1.00
7. Income 4.6 3.35 −.007 −.234 −.041 .133 .092 .162 1.00
8. Gender 1.30 .462 −.049 .059 −.057 .062 .038 .157 .144 1.00
9. Team identity 5.37 1.78 −.080 −.025 .051 .217 .318 .117 .178 −.073 1.00
10. Ephemeral use 2.21 .513 −.084 −.026 −.006 −.235 −.064 .053 −.084 −.062 .035

experience, and 2) the substantive effect of message negativity compose negative eWOM; however, the infused negativity and
and task anxiety on message availability. consumers' anxiety have a negative effect on message life span.
The results of Study 1 provide preliminary evidence to support
H1–H3 within the sports team setting.
Discussion

Study 1 addresses the first three hypotheses related to our Study 2

research questions: How does a negative brand experience
influence consumers' anxiety in the act of creating eWOM?
How do consumers cope with negativity and anxiety when using
ephemeral social media? The results indicate that consumers with
a negative brand experience not only compose social media
messages infused with greater negativity, but they experience
greater anxiety when performing the task compared to those in the
control condition. Furthermore, message negativity and anxiety
during the task significantly influence message availability,
reducing the amount of time the respondents are willing for
their messages to exist on social networks. While it is logical a
negative brand experience would induce greater negativity in a
social media message, it is noteworthy that consumers tend to
limit the longevity of expressed negativity. Consumers still
The purpose of Study 2 is fourfold. First, we seek to replicate
the findings of Study 1 in additional contexts. While we controlled
for team identification with the sports brands in Study 1, the
model's outcomes could be unique to sports fans and negative
team experiences that may not generalize to product or service
brand experiences. For purposes of generalizability, we test the
research model across consumer experiences with six brand
categories inclusive of search goods (clothing, automobiles, and
restaurants) and experience goods (banks, insurance and hotels).
Second, we test whether altering message availability contrib-
utes to a reduction in consumers' task anxiety. Researchers suggest
that studies on negative emotion should not only focus on
in-process emotion but also emotion following a task (Luce 1998).

Fig. 2. Study 1 structural model.

 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59 51

Table 4 perceive that negative eWOM about a non-economic (service)

Direct effects models for message availability (DV).
failure would receive greater scrutiny because such experiences
Model 1 Model 2 Model 3 are more subjective compared to economic (product) failures. For
Controls only Network size + Controls All IVs example, eWOM about non-economic failures may display
Negative experience – – 0.096 discrepancies in cultural capital by signaling differences in
(1.190) knowledge, tastes or practices (Humphreys 2015) or because
Message negativity – – − 0.171 ⁎ some individuals are easily offended or dramatic (Frankowski et
(2.172) al. 2016). In contrast, eWOM concerning economic failures would
Task anxiety – – − 0.156 ⁎
be more objective, fact-based and potentially valuable as an
(1.913)
# Twitter followers – 0.106 0.111 information source for future buyers. Hence, consumers compos-
(1.288) (1.500) ing eWOM about economic experiences may be more likely to
# Facebook friends – 0.169 ⁎ 0.183 ⁎ lengthen message availability as the message increases in
(1.902) (2.036) negativity. However, composing eWOM about more subjective
Team identification 0.108 0.032 0.030
non-economic experiences is likely to raise concerns about ones'
(0.910) (0.285) (0.289)
Age − 0.024 0.031 0.038 online impression. Similarly, message availability is likely to
(0.291) (0.396) (0.465) increase as anxiety increases for negative economic experiences
Gender − 0.036 − 0.064 − 0.056 compared to non-economic experiences due to the objective
(0.481) (0.818) (0.774) nature of a product failure that may be perceived as less
Income − 0.058 − 0.079 − 0.124
controversial.1
(0.735) (1.040) (1.637)
R2 1.7% 5.9% 13.8% H6a. Message availability for economic brand experiences will
⁎ p b .05. be greater than non-economic brand experiences as negativity
increases.
H6b. Message availability for economic brand experiences
A prior study failed to show an expected difference in emotion will be greater than non-economic brand experiences as anxiety
during and post-task (considering an automobile purchase). The increases.
result was attributed to subjects' feeling compelled to keep
emotion consistent (Luce 1998). We suggest that psychological Procedure and Measures
distance (see Trope, Liberman, and Wakslak 2007 for a review)
was an issue due to the hypothetical nature of the task – car We constructed Study 2 using a 2 × 6 × 2 between-subjects
purchasing – among subjects whose purchase intentions for cars factorial design (brand experience: negative or control; product
was unknown. Because our study consists of a task (composing category: clothing, auto, restaurant, bank, insurance or hotel;
eWOM) common to social media users, feelings of anxiety during experience type: economic or non-economic). Subjects from an
the task and post-task would be more salient. Thus, changes in MT pool (n = 232) were randomly assigned to one of the six
consumers' anxiety should be observable and we expect anxiety product categories and each consumer indicated a favorite
will decrease after setting a time limitation on negative eWOM. brand that was piped into one of the scenarios in Table 5
H4. Post-task anxiety will be less than task anxiety. (adapted from Wakefield and Bennett in press). After receiving
the scenario, each subject was instructed to write a Facebook
Third, we provide support for theorizing that impression message they might post regarding the experience. Following
management goals are a source of conflict for consumers sharing the writing task, task anxiety was measured with the same four
negative eWOM. If impression management is an underlying items (Brooks and Schweitzer 2011) in response to the
mechanism, then consumers who increase message availability question, “How did you feel when you were writing that
are likely to use more positive words for self-presentation message about [brand]?” Participants were then asked to select
purposes. a time span for the Facebook message as in Study 1. Then,
H5. The greater the message availability, the greater the post-task anxiety was measured again (Brooks and Schweitzer
self-presentation. 2011) in response to the question, “After using the timer to
self-delete your post about [brand], how do you feel about the
Additionally, the type of negative brand experience, economic message you wrote?”
or non-economic, may also point to impression management as a Control variables included the number of Facebook friends,
relevant goal. Following prior studies (Cambra-Fierro, Melero, gender, age, income, and use of ephemeral systems. In place of
and Sese 2015; Gelbrich and Roschk 2011), economic failures are the team identification control in Study 1, involvement was
due to faulty products while non-economic failures are the result
of poor customer service. Consumers experiencing an economic
1
We do not hypothesize moderation for H1 or the brand experience → message
negativity paths because consumers would feel anxious and negative about a bad
failure are interested in practical gains (e.g., refund or replace-
brand experience whether economic or non-economic and online impression
ment) while those incurring a non-economic failure seek social management goals would not become salient until undertaking the writing task.
gains (e.g., apology, explanation) (Cambra-Fierro, Melero, and Indeed, analysis indicated no moderation along these paths (p N .4; p N .7,
Sese 2015). Relative to impression management, individuals may respectively).
52 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59

Table 5
Brand experience by experience type scenarios.
Experience type
Economic Non-economic
Brand experience Negative You made a purchase with [brand] and the product was You spoke with [brand]'s customer service representative and
really horrible. it was really horrible.
Control You made a purchase with [brand] and the product You spoke with [brand]'s customer service representative and
was really great. it was really great.

measured with a five-item, seven-point bipolar scale (Mittal experienced greater task anxiety and/or greater reduction in
1995) reduced from the original, 20-item set (Zaichkowsky anxiety compared to those choosing longer availability. We used
1985). Discontentment, the only negative consumer emotion multivariate methods to evaluate task anxiety and post-task anxiety
from Richins (1997) not accounted for in the LIWC dictionary by message availability (see Fig. 3). Significant differences were
of negative words used in Study 1, was also included as a found (Wilks Lambda F2, 230 = 4.463, p = .01). Consumers
control and measured with two items (Richins 1997). To designating shorter message availability (b 1 year) had signifi-
measure self-presentation, positive emotion words in each cantly higher levels of task anxiety (M = 2.51) compared to the
message were counted using the LIWC (Pennebaker, Francis, anxiety (M = 1.93) of consumers selecting longer than 1 year
and Booth 2001). The descriptive statistics for the negative and (F1, 231 = 9.969, p = .002). While post-task anxiety did not
control manipulations are in Table 6 and the correlations are in significantly differ based on message availability (Mlow = 1.98,
Table 7. Mhigh = 1.69), an additional analysis indicated that consumers
selecting low availability experienced a significantly greater
Results reduction in anxiety (F1, 231 = 4.809, p = .029). In sum, the use
of ephemeral social media to limit message availability mitigates
Our first objective was to uncover any differences in the consumer anxiety, with greater reduction for those choosing
results based on product category. We performed multiple shorter time spans.
analysis of covariance (MANCOVA) which indicated no Our third objective was to uncover support for impression
significant differences in message negativity or task anxiety by management as a source of conflict for those sharing negative
product category (F = .578, p N .5) or in message availability or eWOM (H5). When composing eWOM about a negative
post-task anxiety by product category (F = 1.5, p N .2). Thus, we experience (n = 121), messages available more than 24 hours
pooled respondent data to assess our second objective — whether (n = 97) are positively correlated with greater use of positive,
anxiety was reduced after limiting message availability. self-presentation words (r = .17, p = .03). Consumers writing
We found support for the idea that ephemeral systems negative eWOM having a longer time span tend to include
contribute to a reduction in consumer anxiety. For all consumers more positive words, which may alleviate the conflict imposed
in the study, task anxiety felt while creating eWOM (M = 2.28) by managing online impressions.
was significantly greater (t = 3.01, p = .003) than post-task Differences in the research model were also found based on
anxiety (M = 1.87) following the use of the ephemeral system, experience type (economic or non-economic). We evaluated a
supporting H4. We also divided the respondents into low and high PLS SEM model for each group as shown in Figs. 4 and 5.
message availability groups: low b 1 year (n = 139) and high For each group, the structural model results mirror the results of
N 1 year (n = 93) to determine if those limiting eWOM availability Study 1 (supporting H1 and H3) with differences in the path
relationship between message negativity and message availability.
Table 6 Fig. 6 shows differences in this path between the economic and
Descriptive statistics by manipulation for Study 2. non-economic groups in support of H6a. As message negativity
Variable Negative condition Control condition increases, eWOM about economic (product) experiences increases
N = 116 N = 116 in message availability at a significantly higher rate than eWOM
Mean SD Mean SD about non-economic (service) experiences (t = 2.09, p = .03). The
experience type did not moderate the relationship between task
Negative emotion words 3.50 2.71 0.627 1.25
Message negativity 0.681 0.691 0.095 0.262 anxiety and message availability (t = −.19, p = .8), failing to
Task anxiety 2.83 1.66 1.70 1.20 support H6b. The control variables discontent and involvement
Post anxiety 2.28 1.55 1.45 1.02 were not significant for either group.
Message availability 4.72 2.17 4.72 1.89
# Facebook friends 322.1 269.3 307.2 289.8
Study 2 Discussion
Age 35.7 9.72 38.6 12.05
Income 4.50 3.06 4.90 3.58
Gender 1.39 0.490 1.43 0.497 The results of Study 2 replicate the findings of Study 1 in terms
Involvement 6.13 1.39 6.36 1.29 of H1 and H3, and demonstrate the generalizability of these
Discontentment 4.07 2.08 1.62 1.26 relationships. Across six product categories, negative brand
Ephemeral use 1.69 0.466 1.67 0.491
experiences resulted in greater anxiety for consumers when
 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59 53

Table 7
Correlation table for Study 2.
Construct 1 2 3 4 5 6 7 8 9 10 11
1. Message neg 1.00
2. Task anxiety .172 1.00
3. Post anxiety .134 .774 1.00
4. Message avail .077 −.252 −.190 1.00
5. # FB friends .019 .144 .170 −.031 1.00
6. Age −.144 −.210 −.189 .058 −.205 1.00
7. Income −.009 .042 .037 .024 .147 .007 1.00
8. Gender −.054 −.023 .021 −.042 −.048 .204 .045 1.00
9. Ephemeral use .124 −.059 −.074 .095 −.226 .302 −.094 .136 1.00
10. Involvement −.133 −.030 .016 .065 .101 .183 .105 .083 .050 1.00
11. Discontent .368 .598 .468 −.145 .007 −.221 .098 −.067 −.003 −.093 1.00

creating eWOM and led to limiting message availability. This
finding reinforces the idea that sharing about a poor brand
experience in social media poses a threat to the consumer. To cope
with the threat, the consumer places greater limitations on the
message's online availability and availability decreases as more
negative wording is included in the message.
A key finding of Study 2 is a significant reduction in consumer
anxiety occurs after utilizing the self-deleting feature of an
ephemeral system to curtail message availability. If anxiety is a
motivator for an individual to limit message availability, then
anxiety should decrease following use of an ephemeral system.
Not only did anxiety decrease after restricting message availability,
but those who felt significantly greater anxiety while creating
eWOM also had the greatest reduction in anxiety after using the
timer. Consumers using ephemeral social media experienced a
significant reduction of anxiety when sharing negative brand
experiences.
Prior to Study 2, an explanation for consumer anxiety in
creating negative eWOM was not proffered. The results of Study
2 provide evidence that negative eWOM is a threat to online
impression management. For example, results show a) the use of
positive emotion words by those who increased message
availability when sharing about a negative brand experience,
and b) consumers limiting message availability more for
non-economic (service) experiences. This implies two means
through which consumers appear to cope with the conflict arising
when sharing negative brand experiences in eWOM and
3
2.5
Anxiety
maintaining a positive online image — incorporate positive
emotion words into the negative message or decrease message
availability. Consumers not including positivity in their messages
were reluctant to increase message availability. Furthermore,
consumers creating eWOM based on the economic or non-
economic brand experience preferred significantly different
message availabilities. This result lends credence to impression
management as an underlying concern. Consumers composing
negative eWOM about an economic experience (e.g., poor
product) tend to increase the availability of the social media
message at a greater rate as more negativity is used, compared to
eWOM about a non-economic experience (e.g., poor service). It is
likely consumers believe a negative service evaluation would
attract more scrutiny from their social media recipients due to its
subjectivity. In contrast, negative economic (product) eWOM
would be more objective because the focus would be on a
verifiable, unbiased product experience. For illustrative purposes,
consider the noticeable subjectivity in the following qualitative
responses from those sharing a negative, non-economic (service)
brand experience:
a. Wow. Worst experience I ever had at Toyota. Sales rep was
cold, had no time for me and couldn't answer even the most
basic questions. Plus his hair was very odd looking and
distracting. Must have been new hire.
b. Talk about not caring, Chipotle “non customer service” was
snarky, but not even interesting snarky, more like bad
hipster sarcasm. Hipsters are so over, btw, Chipotle. 1999
called and they don't want you back anyway.
In contrast, the qualitative responses from those sharing a
negative, economic (product) brand experience are notably
more objective and helpful for consumers who may consider a
future purchase:

Low Availability
2 High Availability
a. Really can't believe this. I bought a GAP sweater and after
two wears I noticed some of the fabric coming apart. I
1.5 washed it and it's all completely torn up. I might as well
have thrown my money out. This is crap.
b. I went to In-N-Out [Burger] today and it wasn't great.
1
Task Post-Task Ordered a regular with grilled onions and cheese and a
order of fries. They gave me a burger, no grilled onions. I
Fig. 3. Task anxiety and post-task anxiety by low and high message availability. also had to wait over 10 minutes for my fries. Terrible.
54 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59

Fig. 4. Study 2 structural model for economic (product) brand experience (N = 116).

The experience type did not moderate the relationship between
task anxiety and message availability (H6b). This implies that
consumers do not alter message availability by experience type
when they feel more/less anxious, but they do shift the availability
by experience type when the content of the message is more
negative (H6a). Consumers may feel greater potential for scrutiny
when sharing non-economic experiences; however, the behavioral
response (message availability) relative to experience type is
activated with message negativity.
Implications
The effects of electronic word-of-mouth (eWOM) have long
captured the interests of brand marketers because they affect key
performance indicators such as sales (Duan, Gu, and Whinston
2008), customer value and loyalty (Gruen, Osmonbekov, and
Czaplewski 2006), and consumer recommendations (Bickart and
Schindler 2001). While eWOM studies have brought attention
to consumers' feelings antecedent to creating eWOM, consumer
emotion while in the process of writing eWOM has been
overlooked. This oversight is an opportunity to contribute to the
knowledgebase of consumer behavior surrounding emotion in an
action performed by a large proportion of consumers (i.e., creating
eWOM). Across two studies, we found that consumers composing
eWOM about a negative brand encounter experience anxiety, and
the anxiety, as suggested in Study 2, relates to online impression
management. As the brand experience stimulates anxiety (and
negativity) while messages are being composed, consumer anxiety
leads to the altering of message availability. Thus, we offer insights
into why and how eWOM can be constrained. We found these
results robust for a variety of brand categories (sports, search and
experience goods) as well as economic and non-economic brand

Fig. 5. Study 2 structural model for non-economic (service) brand experience (N = 116).
 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59
Message Availability
Economic
Non-Economic
Message Negativity
Fig. 6. Message availability of economic and non-economic experiences by
message negativity.
experience types. Given the importance of eWOM and the
widespread use of social media, understanding this process is
valuable for brand managers for several reasons.
First, webcare management is an ongoing concern of brands
that actively monitor online brand sentiment. While a responsive,
proactive webcare strategy within a consumer platform has been
cautioned (Fournier and Avery 2011; Van Noort and Willemsen
2012), we suggest that a reactive strategy may also be misguided,
particularly when a consumer is experiencing anxiety and
negativity related to a brand experience. Public redress may seem
an appropriate response to negative eWOM because a wider
audience observing a positive resolution could lead to positive
WOM (Sernovitz 2012). However, doing so may extend the
availability (likely to a larger audience) of the negative eWOM,
draw greater online attention to the negative brand experience, and
prolong the incident for the consumer. If a negative experience is
worsened by the brand's response, then the resolution would be of
little benefit to the consumer or the brand. Many consumers cope
with highly emotional experiences by avoiding further action
(Duhachek 2005; Luce 1998; Stephens and Gwinner 1998); thus, a
brand's reaction may discourage the consumer from sharing
negative brand experiences in the future. In light of our findings,
the better strategy may be to address anxious and negative
consumers privately and/or offline. This alternative is less likely to
compound or prolong a negative emotion-related brand experience.
The consumer would satisfy the need to share an emotional
experience and the brand would have knowledge of the negative
experience. If the consumer chooses to share the resolution online,
all the better.
Additionally, we argue impression management is a salient
concern of consumers sharing negative eWOM. This concern
leads to anxiety and consumers' coping via the use of positive
words or altering message availability. Interestingly, the type of
brand experience (economic or non-economic) has an effect on
message availability. We infer from consumers' message
availability decisions that negativity about economic (product)
brand experiences is less threatening to impression management
because it has a more objective basis. Hence, negative eWOM
about economic brand encounters will tend toward greater online
longevity. For example, recipients of the messages shown in the
excerpts from Study 2 have no reason not to believe that Gap and
In-N-Out are responsible for the consumers' negative experiences
55
and thus may rationally infer that future customers would
experience the same. Message recipients may even appreciate
negative eWOM of economic brand experiences because
consumer reviews are seen as a fundamental information source
for future purchases (Mathwick and Mosteller 2017). In contrast,
negative eWOM about non-economic (service) experiences are
potentially subject to greater scrutiny, and a mindful consumer
may limit message availability to minimize the threat to a positive
online impression. Brand managers of non-economic offerings
should be aware that negative online sentiment may be available
for a shorter time and/or diluted with positive words to accom-
modate impression management concerns. Hence, consumers'
negative non-economic experiences could be much worse than
what is shared online. Additional investigation, preferably offline,
would be needed for managers to truly understand the gravity of a
non-economic grievance expressed in social media.
Scholars have pointed out that ‘the venue matters’ when
observing consumer behavior in social media as consumer
actions are not equivocal across platforms (viz. Hanna, Rohm,
and Crittenden 2011; Schweidel and Moe 2014). We find in an
ephemeral social media context, consumers will limit/lengthen
the lifespan of messages based on a brand experience, the level of
anxiety and the amount of negativity in the eWOM. Altering
the availability of messages gives consumers control over
eWOM (Wakefield and Bennett in press) and our study suggests
consumers may use this control to alleviate feelings of anxiety
that arise while composing eWOM and/or managing online
impressions. An emotional trade-off occurs when sacrificing
one goal for another. Interestingly, we find the trade-off
(i.e., impression management over negative eWOM creation, or
vice-versa) is either removed or mitigated when the consumer
controls message availability. With lower anxiety about the
negative eWOM, there would be less focus on managing online
impressions and consumers may be more inclined to share
negative eWOM. This is not only helpful for consumers, but
also for brands because ephemeral social media is generally
challenging to evaluate (Piwek and Joinson 2016). Ephemeral
systems provide a way for consumers to cope with negative
brand experiences through the sharing of sentiment, rather than
avoidance, and they give brands an opportunity to respond.
Brands should also be cognizant of the timing aspects of
eWOM shared in ephemeral social media. While services exist to
archive eWOM and distribute them throughout an organization
(e.g., HelpSocial.com), brands desiring to respond to consumer
issues directly may have a limited time to do so. We found the most
negative messages sent by the most anxious consumers were
available for the shortest time. Arguably, the most upset customers
would require attention sooner if the problem is to be resolved.
From another point of view, ephemeral social media may be an
effective venue to solicit negative eWOM from consumers. Brands
currently ask customers to share their concerns directly with the
organization using specific channels such as a customer service
hotline. However, a considerable proportion of consumers already
use social media to voice complaints (Morrison 2016). Hence,
consumers may not need much incentive to choose ephemeral
social media to voice sentiment, and brands could develop
resolution strategies appropriate for an ephemeral platform.
56 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59
Additionally, brands could use ephemeral systems to thwart
the detrimental effects of expected negative brand experiences.
Recall our sports scenario in which a professional team incurred a
devastating loss. Teams or leagues are generally aware in
real-time when fans are having a poor brand experience or are
likely to have one (i.e., one team is far outmatched), and in such
instances, promoting ephemeral messaging use would be
beneficial. For example, a sports team might increase marketing
promotions in Snapchat rather than Twitter during a game in
which they are losing horribly. One purpose may be to encourage
fans' negative eWOM in a short-lived atmosphere rather than a
social media platform conducive to prolonging the negative
experience. The same approach may be applied to negative
product and service experiences outside of sport. Accommodat-
ing and encouraging eWOM in an ephemeral system may be
beneficial to both consumers and brands following a product
failure or recall. While brands generally have no say over where
consumers place their negative eWOM, promoting the ‘control’
aspect of ephemeral systems to consumers might incentivize use
and brands could provide resolution in a managed environment.
Future Research
There are multiple directions of study for scholars interested in
this research stream. Our focus on anxiety in eWOM differs from
prior studies of consumer anxiety, such as those focusing on
consumers' math anxiety in relation to pricing (Feng, Suri, and
Bell 2014; Suri, Monroe, and Koc 2013). Given the evidence of
anxiety in social media use (Andreassen et al. 2016) and the
usefulness of social media to marketers (e.g., De Vries, Gensler,
and Leeflang 2012), examining additional behaviors related to
anxiety in social media eWOM may be a fruitful path. For
example, researchers might focus on the potential for anxiety to
form in recipients (rather than sharers) of negative eWOM to
determine how their anxiety might influence brand attitudes as
well as online responses to the sharer and the brand. Does a
sharer's negative eWOM motivate negative eWOM from the
recipient? How do recipients of negative eWOM consider their
online impression and their audience in composing a response?
Do recipients of negative eWOM attribute the negativity to the
brand, to the sharer, or both? These are interesting and important
questions in the quest to fully understand how consumers'
brand-related emotions influence attitudes and online behavior.
The prediction of consumer anxiety in the creation of eWOM is
an important topic that should be broadened to include other
consumer emotions. Current studies generally categorize brand
sentiment as positive or negative (Fay and Larkin 2017; Ma, Sun,
and Kekre 2015; Zhang, Bhattacharyya, and Ram 2016).
However, interesting effects and insights are likely to be gleaned
from observing, measuring and examining other specific consumer
emotions (see Richins 1997) with similar methods. For example,
positive brand experiences may elicit excitement, optimism or joy
in the creation of eWOM that influence how and to whom the
experience is shared, how the eWOM is received, and how long the
eWOM is likely to exist online. Investigations into these emotions
should consider using the Linguistic Inquiry and Word Count
(LIWC) (Pennebaker, Francis, and Booth 2001) text analysis
program to compute the frequency of expressed emotion and
compare objective measures of emotion to reported feelings and
behaviors. Studies that broadly categorize brand sentiment as
positive or negative may also benefit from using the LIWC as a
manipulation check. Brand managers and their webcare strategies
would benefit from a more comprehensive investigation of specific
consumer emotions related to brand experiences and eWOM. Web
services such as Google now offer consumer sentiment analyses
that include the likelihood of several emotions including anger,
sadness, and joy (Casalboni 2016). Academic research is currently
lagging in this area.
While we find that shortening message availability can
alleviate anxiety, there may be other ways that consumers may
reduce such feelings. As some complain to receive social
support (Hennig-Thurau et al. 2004), what are the effects of
receiving social support? Perhaps consumers who receive
social support feel less anxiety, experience enhanced social
status and are emboldened to share negative sentiment in the
future. For example, a sports fan sharing negative comments on
Facebook after a team's poor performance may target other
disappointed fans who can reciprocate those feelings and
provide empathy and appreciation through likes, comments and
shares. Future research should consider how sharing negative
sentiment may contribute to impression management.
Finally, the issue of control over messages by consumers and
brands should be studied further. Wakefield and Bennett (in press)
suggest that ephemeral social media give consumers greater control
over online messages than ever before; however, consumers' goals
relative to control are unknown. We find that control over message
availability may reduce consumer anxiety generated when creating
eWOM. However, brands generally have little control over
consumers' online sentiment. Yet, brands could exert some control
over the placement and longevity of messages by leveraging
ephemeral system use in certain settings with incentives and
promotions. Just as brands often encourage customers to contact
customer service representatives through particular communication
methods (e.g., phone hotline), future research might examine how
brands achieve control through promoting particular types of social
media for sharing certain types of feedback.
Limitations
Common method variance (CMV) is a potential statistical
artifact that may result in significant relationships in the model
due to systematic bias. We randomized the order of the survey
items between subjects in an effort to reduce CMV. Additionally,
we applied Harman's single-factor test (Podsakoff et al. 2003) to
test for excessive bias. We performed confirmatory factor
analysis for the three main constructs which resulted in the
extraction of three factors. The first factor accounted for 38% of
the covariance suggesting that CMV is limited because multiple
factors emerged with no single factor accounting for a majority of
the covariance. The self-report methodology is also a limitation
that may be overcome by lab studies that incorporate control
groups and a physiological measure of anxiety. Additional
manipulations of the brand experience would also strengthen our
findings.
 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59
Despite limitations, the present study has several strengths.
First, we incorporated a manipulation to focus on the organismic
responses in S–O–R theory that occur when an individual is
exposed to a stimulus. We successfully manipulated the brand
experience in Study 1 providing evidence that the negative brand
scenario was the impetus for message negativity and task anxiety.
Study 2 included different brand categories to highlight consumer
anxiety and negative eWOM in various contexts as well as
provided evidence for the presence of impression management
when creating negative eWOM. Additionally, the sampled
respondents were diverse in terms of age implying different life
stages that increases the generalizability of the results to a broad
range of social media users. However, it is unknown whether the
model results are applicable to a different culture of social media
users.
References
Adler, Patricia A. and Peter Adler (2008), “The Cyber Worlds of Self-injurers:
Deviant Communities, Relationships, and Selves,” Symbolic Interaction,
31, 1, 33–56.
Anderson, Norman H. (1965), “Averaging Versus Adding as a Stimulus-
combination Rule in Impression Formation,” Journal of Experimental
Psychology, 70, 4, 394–400.
Andreassen, Cecilie Schou, Joël Billieux, Mark D. Griffiths, Daria J. Kuss,
Zsolt Demetrovics, Elvis Mazzoni, and Ståle Pallesen (2016), “The
Relationship Between Addictive Use of Social Media and Video Games
and Symptoms of Psychiatric Disorders: A Large-scale Cross-sectional
Study,” Psychology of Addictive Behaviors, 30, 2, 252–62.
Barasch, Alixandra and Jonah Berger (2014), “Broadcasting and Narrowcasting:
How Audience Size Affects What People Share,” Journal of Marketing
Research, 51, 3, 286–99.
Batra, Rajeev, Aaron Ahuvia, and Richard P. Bagozzi (2012), “Brand Love,”
Journal of Marketing, 76, 2, 1–16.
Bayer, Joseph B., Nicole B. Ellison, Sarita Y. Schoenebeck, and Emily B. Falk
(2015), “Sharing the Small Moments: Ephemeral Social Interaction on
Snapchat,” Information, Communication & Society, 19, 7, 956–77.
Berger, Jonah (2014), “Word of Mouth and Interpersonal Communication: A
Review and Directions for Future Research,” Journal of Consumer
Psychology, 24, 4, 586–607.
Berry, Leonard L. (2005), “Relationship Marketing of Services — Growing
Interest, Emerging Perspectives,” Journal of the Academy of Marketing
Science, 23, 4, 236–45.
Bickart, Barbara and Robert M. Schindler (2001), “Internet Forums as
Influential Sources of Consumer Information,” Journal of Interactive
Marketing, 15, 3, 31–40.
Blodgett, Jeffrey G., Kirk L. Wakefield, and James H. Barnes (1995), “The
Effects of Customer Service on Consumer Complaining Behavior,” Journal
of Services Marketing, 9, 4, 31–42.
Brakus, J. Joško, Bernd H. Schmitt, and Lia Zarantonello (2009), “Brand
Experience: What Is it? How Is it Measured? Does it Affect Loyalty?”
Journal of Marketing, 73, 3, 52–68.
Brooks, Alison Wood and Maurice E. Schweitzer (2011), “Can ‘Nervous Nelly’
Negotiate? How Anxiety Causes Negotiators to Make Low First Offers,
Exit Early, & Earn Less Profit,” Organizational Behavior and Human
Decision Processes, 115, 1, 43–54.
Burchard, Jeremy (2017), “Brad Paisley Makes Fun of Online Culture in Hilarious
New Video for ‘Selfie#Theinternetisforever’,” (April 4) available at http://
www.wideopencountry.com/brad-paisley-selfietheinternetisforever/.
Cambra-Fierro, Jesus, Iguacel Melero, and F. Javier Sese (2015), “Managing
Complaints to Improve Customer Profitability,” Journal of Retailing, 91, 1,
109–24.
Caplan, Scott E. (2007), “Relations among Loneliness, Social Anxiety, and
Problematic Internet Use,” Cyberpsychology & Behavior, 10, 2, 234–42.
57
Casalboni, Alex (2016), “Build Powerful Applications that See and Understand
the Content of Images with the Google Vision API,” (January 28) available at
https://cloudacademy.com/blog/google-vision-api-image-analysis/.
Chandler, Jennifer D. and Robert F. Lusch (2015), “Service Systems: A
Broadened Framework and Research Agenda on Value Propositions,
Engagement, and Service Experience,” Journal of Service Research, 18,
1, 6–22.
Chang, Hyo-Jung, Molly Eckman, and Ruoh-Nan Yan (2011), “Application of the
Stimulus–Organism–Response Model to the Retail Environment: The Role of
Hedonic Motivation in Impulse Buying Behavior,” The International Review
of Retail, Distribution and Consumer Research, 21, 3, 233–49.
Chebat, Jean-Charles, Moshe Davidow, and Isabelle Codjovi (2005), “Silent
Voices: Why Some Dissatisfied Consumers Fail to Complain,” Journal of
Service Research, 7, 4, 328–42.
Cohen, Jacob (1988), Statistical power analysis: A computer program. Hillsdale,
NJ: Routledge.
——— (1992), “A Power Primer,” Psychological Bulletin, 112, 1, 155–9.
Correa, Teresa, Amber Willard Hinsley, and Homero Gil De Zuniga (2010),
“Who Interacts on the Web? The Intersection of Users' Personality and
Social Media Use,” Computers in Human Behavior, 26, 2, 247–53.
Coupland, Jennifer Chang (2005), “Invisible Brands: An Ethnography of
Households and the Brands in their Kitchen Pantries,” Journal of Consumer
Research, 32, 1, 106–18.
De Vries, Lisette, Sonja Gensler, and Peter S.H. Leeflang (2012), “Popularity of
Brand Posts on Brand Fan Pages: An Investigation of the Effects of Social
Media Marketing,” Journal of Interactive Marketing, 26, 2, 83–91.
Duan, Wenjing, Bin Gu, and Andrew B. Whinston (2008), “Do Online Reviews
Matter? An Empirical Investigation of Panel Data,” Decision Support
Systems, 45, 4, 1007–16.
Dubois, David, Andrea Bonezzi, and Matteo De Angelis (2016), “Sharing with
Friends Versus Strangers: How Interpersonal Closeness Influences Word-
of-Mouth Valence,” Journal of Marketing Research, 53, 5, 712–27.
Duhachek, Adam (2005), “Coping: A Multidimensional, Hierarchical Frame-
work of Responses to Stressful Consumption Episodes,” Journal of
Consumer Research, 32, 1, 41–53.
Ellison, Nicole B., Rebecca Heino, and Jennifer Gibbs (2006), “Managing
Impressions Online: Self-presentation Processes in the Online Dating Environ-
ment,” Journal of Computer-Mediated Communication, 11, 2, 415–41.
———, Jessica Vitak, Charles Steinfield, Rebecca Gray, and C. Cliff Lampe
(2011), Negotiating Privacy Concerns and Social Capital Needs in a Social
Media Environment. Berlin: Springer19–32.
Eroglu, Sevgin A., Karen A. Machleit, and Lenita M. Davis (2001),
“Atmospheric Qualities of Online Retailing: A Conceptual Model and
Implications,” Journal of Business Research, 54, 2, 177–84.
Eversley, Melanie (2011), “Library of Congress to Store Tweets Based
on Twitter Deal,” (December 7) available at http://content.usatoday.com/
communities/ondeadline/post/2011/12/library-of-congress-to-store-tweets-
based-on-twitter-deal/1#.WIYRBBsrKUk.
Fay, Brad and Rick Larkin (2017), “Why Online Word-of-Mouth Measures
Cannot Predict Brand Outcomes Offline,” Journal of Advertising Research,
57, 2, 132–43.
Feng, Shan, Rajneesh Suri, and Monique Bell (2014), “Does Classical Music
Relieve Math Anxiety? Role of Tempo on Price Computation Avoidance,”
Psychology and Marketing, 31, 7, 489–99.
Fisher, Robert and Kirk Wakefield (1998), “Factors Leading to Group
Identification: A Field Study of Winners and Losers,” Psychology and
Marketing, 15, 1, 23–40.
Fiske, Susan T. (2001), “Social and Societal Pragmatism: Commentary on
Augustinos, Gaskell, and Lorenzi-Cioldi,” in Representations of the Social:
Bridging Research Traditions. Kay Deaux, Gina Philogene, editors. New York:
Blackwell. 249–53.
Fornell, Claes and David F. Larcker (1981), “Structural Equation Models with
Unobservable Variables and Measurement Error: Algebra and Statistics,”
Journal of Marketing Research, 18, 3, 382–8.
Fournier, Susan and Jill Avery (2011), “The Uninvited Brand,” Business
Horizons, 54, 3, 193–207.
Frankowski, Scott, Amber K. Lupo, Brandt A. Smith, Mosi Dane'El, Corin
Ramos, and Osvaldo F. Morera (2016), “Developing and Testing a Scale to
58 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59
Measure Need for Drama,” Personality and Individual Differences, 89,
192–201.
Gelbrich, Katja and Holger Roschk (2011), “A Meta-analysis of Organizational
Complaint Handling and Customer Responses,” Journal of Service
Research, 14, 1, 24–43.
Gruen, Thomas W., Talai Osmonbekov, and Andrew J. Czaplewski (2006),
“eWOM: The Impact of Customer-to-Customer Online Know-How
Exchange on Customer Value and Loyalty,” Journal of Business Research,
59, 4, 449–56.
Hair Jr., Joseph F., William C. Black, Barry J. Babin, and Rolph E. Anderson
(2010), Multivariate Data Analysis. Englewood Cliffs, NJ: Prentice Hall.
———, G. Tomas, M. Hult, Christian Ringle, and Marko Sarstedt (2017), A
Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM).
Thousand Oaks, CA: Sage Publications.
Hancock, Jeffrey T. and Philip J. Dunham (2001), “Impression Formation
in Computer-mediated Communication Revisited: An Analysis of the
Breadth and Intensity of Impressions,” Communication Research, 28, 3,
325–47.
Hanna, Richard, Andrew Rohm, and Victoria L. Crittenden (2011), “We're All
Connected: The Power of the Social Media Ecosystem,” Business Horizons,
54, 3, 265–73.
Hartung, Adam (2016), “Why Is Snapchat Worth $20B? The Value of
Implementing Trends,” (May 27) available at http://www.forbes.com/
sites/adamhartung/2016/05/27/why-is-snapchat-worth-20b-the-value-of-
implementing-trends/#309d003153fa.
Hasinoff, Amy Adele (2013), “Sexting as Media Production: Rethinking Social
Media and Sexuality,” New Media & Society, 15, 4, 449–65.
Hennig-Thurau, Thorsten, Kevin P. Gwinner, Gianfranco Walsh, and Dwayne
D. Gremler (2004), “Electronic Word-of-Mouth via Consumer-Opinion
Platforms: What Motivates Consumers to Articulate Themselves on the
Internet?” Journal of Interactive Marketing, 18, 1, 38–52.
Humphreys, Ashlee (2015), Social Media: Enduring Principles. Oxford
University Press.
Kahn, Jeffrey H., Renee M. Tobin, Audra E. Massey, and Jennifer A. Anderson
(2007), “Measuring Emotional Expression with the Linguistic Inquiry and
Word Count,” The American Journal of Psychology, 120, 2, 263–86.
Kim, Jiyoung and Sharron J. Lennon (2013), “Effects of Reputation and
Website Quality on Online Consumers' Emotion, Perceived Risk and
Purchase Intention: Based on the Stimulus–Organism–Response Model,”
Journal of Research in Interactive Marketing, 7, 1, 33–56.
Kotler, Philip (1997), Marketing Management 7th edition. Upper Saddle River,
NJ: Prentice Hall.
Krämer, Nicole C. and Stephan Winter (2008), “Impression Management 2.0:
The Relationship of Self-esteem, Extraversion, Self-efficacy, and Self-
presentation within Social Networking Sites,” Journal of Media Psychol-
ogy, 20, 3, 106–16.
Kross, Ethan, Philippe Verduyn, Emre Demiralp, Jiyoung Park, David Seungjae
Lee, Natalie Lin, Holly Shablack, John Jonides, and Oscar Ybarra (2013),
Facebook Use Predicts Declines in Subjective Well-being in Young Adults,
https://doi.org/10.1371/journal.pone.0069841 (August 14).
Lazarus, Richard S. and Susan Folkman (1984), Stress, Appraisal, and Coping.
New York: Springer.
——— (1991), Emotion and Adaptation. New York: Oxford University Press.
Lee, Bianca W. and Lexine A. Stapinski (2012), “Seeking Safety on the
Internet: Relationship between Social Anxiety and Problematic Internet
Use,” Journal of Anxiety Disorders, 26, 1, 197–205.
Luce, Mary Frances (1998), “Choosing to Avoid: Coping with Negatively
Emotion-laden Consumer Decisions,” Journal of Consumer Research, 24,
4, 409–33.
———, James R. Bettman, and John W. Payne (2001), “Emotional Decisions:
Tradeoff Difficulty and Coping in Consumer Choice,” Monographs of the
Journal of Consumer Research, 1, 1–209.
Ma, Liye, Baohong Sun, and Sunder Kekre (2015), “The Squeaky Wheel Gets
the Grease — An Empirical Analysis of Customer Voice and Firm
Intervention on Twitter,” Marketing Science, 34, 5, 627–45.
Madrigal, Alexis (2013), “71% of Facebook Users Engage in ‘Self-censorship’,”
(April 15) available at https://www.theatlantic.com/technology/archive/2013/
04/71-of-facebook-users-engage-in-self-censorship/274982/.
Mankad, Aditi, Sandy Gordon, and Karen Wallman (2009), “Psycholinguistic
Analysis of Emotional Disclosure: A Case Study in Sport Injury,” Journal
of Clinical Sport Psychology, 3, 2, 182–96.
Manovich, Lev (2009), “The Practice of Everyday (Media) Life: From Mass
Consumption to Mass Cultural Production?” Critical Inquiry, 35, 2,
319–31.
Marwick, Alice E. and Danah Boyd (2011), “I Tweet Honestly, I Tweet
Passionately: Twitter Users, Context Collapse and the Imagined Audience,”
New Media & Society, 13, 1, 114–33.
Mathwick, Charla and Jill Mosteller (2017), “Online Reviewer Engagement: A
Typology Based on Reviewer Motivations,” Journal of Service Research,
20, 2, 204–18.
McGraw, A. Peter, Caleb Warren, and Christina Kan (2014), “Humorous
Complaining,” Journal of Consumer Research, 41, 5, 1153–71.
Mehrabian, Albert and James A. Russell (1974), An Approach to Environmen-
tal Psychology. Cambridge, MA: MIT Press.
Mittal, Banwari (1995), “A Comparative Analysis of Four Scales of Consumer
Involvement,” Psychology and Marketing, 12, 7, 663–82.
Morrison, Kimberlee (2016), “80% of Consumers Use Digital Devices to Contact
Customer Service (Report),” (August 17) available at http://www.adweek.
com/digital/80-of-consumers-use-digital-devices-to-contact-customer-service-
report/.
Newman, Mark W., Debra Lauterbach, Sean A. Munson, Paul Resnick, and
Margaret E. Morris (2011), “It's Not that I Don't Have Problems, I'm
Just Not Putting Them on Facebook: Challenges and Opportunities in
Using Online Social Networks for Health,” Proceedings of the ACM 2011
Conference on Computer Supported Cooperative Work. 341–50.
Newport, Frank (2016), “American Public Opinion, Terrorism and Guns,”
(June 13) available at http://www.gallup.com/opinion/polling-matters/
192695/american-public-opinion-terrorism-guns.aspx.
Nunnally, Jum (1978), Psychometric Methods. New York: McGraw-Hill.
O'Cass, Aron and Liem Viet Ngo (2007), “Market Orientation Versus
Innovative Culture: Two Routes to Superior Brand Performance,” European
Journal of Marketing, 41, 7/8, 868–87.
Peck, Joann and Terry L. Childers (2003), “To Have and to Hold: The Influence of
Haptic Information on Product Judgments,” Journal of Marketing, 67, 2, 35–48.
Pennebaker, James W. (1993), “Mechanisms of Social Constraint,” in
Handbook of Mental Control. D.M. Wegner, J.W. Pennebaker, editors.
Englewood Cliffs, NJ: Prentice Hall. 200–19.
———, Emmanuelle Zech, and Bernard Rimé (2001), “Disclosing and
Sharing Emotion: Psychological, Social, and Health Consequences,”
Handbook of Bereavement Research: Consequences, Coping, and Care,
517–43.
———, Martha E. Francis, and Roger J. Booth (2001), Linguistic Inquiry and
Word Count: LIWC. Mahwah, NJ: Lawrence Erlbaum Associates.
Piwek, Lukasz and Adam Joinson (2016), ““What Do They Snapchat About?”
Patterns of Use in Time-limited Instant Messaging Service,” Computers in
Human Behavior, 54, 1, 358–67.
Podsakoff, Philip M., Scott B. MacKenzie, Jeong-Yeon Lee, and Nathan P.
Podsakoff (2003), “Common Method Biases in Behavioral Research: A
Critical Review of the Literature and Recommended Remedies,” Journal of
Applied Psychology, 88, 5, 879–903.
Poltash, Nicole A. (2012), “Snapchat and Sexting: A Snapshot of Baring your
Bare Essentials,” Richmond Journal of Law and Technology, 19, 4, 1–24.
Richard, Marie-Odile and Ramdas Chandra (2005), “A Model of Consumer
Web Navigational Behavior: Conceptual Development and Application,”
Journal of Business Research, 58, 8, 1019–29.
Richins, Marsha L. (1997), “Measuring Emotions in the Consumption
Experience,” Journal of Consumer Research, 24, 2, 127–46.
Ringle, Christian M., Sven Wende, and Jan-Michael Becker (2015), SmartPLS 3.
Boenningstedt: SmartPLS GmbH (available at http://www.smartpls.com).
Roesner, Franziska, Brian T. Gill, and Tadayoshi Kohno (2014), “Sex, Lies, or
Kittens? Investigating the Use of Snapchat's Self-destructing Messages,” in
International Conference on Financial Cryptography and Data Security,
Berlin Heidelberg: Springer, 64–76.
Rosen, Jeffrey (2010), “The Web Means the End of Forgetting,” (July 21)
available at http://www.nytimes.com/2010/07/25/magazine/25privacy-t2.
html?pagewanted=all.
 L.T. Wakefield, R.L. Wakefield / Journal of Interactive Marketing 41 (2018) 44–59
Russell, Jon (2017), “Snow, the Asian Snapchat Clone Facebook Tried to Buy,
Claims 40–50M Monthly Users,” (January 4) available at https://techcrunch.
com/2017/01/04/snow-the-asian-snapchat-clone-facebook-tried-to-buy-claims-
40-50m-monthly-users/.
Schwartz, Gary E. (1990), “Psychobiology of Repression and Health: A Systems
Approach,” Repression and Dissociation: Implications for Personality
Theory, Psychopathology, and Health, 405–34.
Schweidel, David A. and Wendy W. Moe (2014), “Listening in on Social
Media: A Joint Model of Sentiment and Venue Format Choice,” Journal of
Marketing Research, 51, 4, 387–402.
Sernovitz, Andy (2012), “10 Ways to Deal with Upset Customers Using Social
Media,” (February 22) available at http://www.socialmediaexaminer.com/
10-ways-to-deal-with-upset-customers-using-social-media/.
Singh, Val, Savita Kumra, and Susan Vinnicombe (2002), “Gender and
Impression Management: Playing the Promotion Game,” Journal of Business
Ethics, 37, 1, 77–89.
Sparks, Beverley A. and Victoria Browning (2010), “Complaining in
Cyberspace: The Motives and Forms of Hotel Guests' Complaints Online,”
Journal of Hospitality Marketing & Management, 19, 7, 797–818.
Statista (2015), “Statistics and Facts about Social Media Marketing in the
United States,” (May 4) available at http://www.statista.com/topics/1538/
social-media-marketing/.
Steelman, Zachary R., Bryan I. Hammer, and Moez Limayem (2014), “Data
Collection in the Digital Age: Innovative Alternatives to Student Samples,”
MIS Quarterly, 38, 2, 355–78.
Stephens, Nancy and Kevin P. Gwinner (1998), “Why Don't Some People
Complain? A Cognitive-Emotive Process Model of Consumer Complaint
Behavior,” Journal of the Academy of Marketing Science, 26, 3, 172–89.
Stieglitz, Stefan and Linh Dang-Xuan (2013), “Emotions and Information
Diffusion in Social Media — Sentiment of Microblogs and Sharing
Behavior,” Journal of Management Information Systems, 29, 4, 217–47.
Stone, Mervyn (1974), “Cross-validatory Choice and Assessment of Statistical
Predictions,” Journal of the Royal Statistical Society, 36, 2, 111–47.
Suler, John (2004), “The Online Disinhibition Effect,” Cyberpsychology &
Behavior, 7, 3, 321–6.
Suri, Rajneesh, Kent B. Monroe, and Umit Koc (2013), “Math Anxiety and its
Effects on Consumers' Preference for Price Promotion Formats,” Journal of
the Academy of Marketing Science, 41, 3, 271–82.
59
Tausczik, Yla R. and James W. Pennebaker (2010), “The Psychological
Meaning of Words: LIWC and Computerized Text Analysis Methods,”
Journal of Language and Social Psychology, 29, 1, 24–54.
Trope, Yaacov, Nira Liberman, and Cheryl Wakslak (2007), “Construal Levels and
Psychological Distance: Effects on Representation, Prediction, Evaluation, and
Behavior,” Journal of Consumer Psychology, 17, 2, 83–95.
Van Doorn, Jenny, Katherine N. Lemon, Vikas Mittal, Stephan Nass, Doreén
Pick, Peter Pirner, and Peter C. Verhoef (2010), “Customer Engagement
Behavior: Theoretical Foundations and Research Directions,” Journal of
Service Research, 13, 3, 253–66.
VanMeter, Rebecca A., Douglas B. Grisaffe, and Lawrence B. Chonko (2015),
“Of “Likes” and “Pins”: The Effects of Consumers' Attachment to Social
Media,” Journal of Interactive Marketing, 32, 4, 70–88.
Van Noort, Guda and Lotte M. Willemsen (2012), “Online Damage Control:
The Effects of Proactive Versus Reactive Webcare Interventions in
Consumer-generated and Brand-generated Platforms,” Journal of Interac-
tive Marketing, 26, 3, 131–40.
Wagner, Kurt (2017), “Facebook Copied Snapchat a Fourth Time, and Now All
its Apps Look the Same,” (March 28) available at https://www.recode.net/
2017/3/28/15079774/facebook-stories-snapchat-instagram-copy.
Wakefield, Lane T. and Gregg Bennett (in press), “Sports Fan Experience:
Electronic-Word-of-Mouth in Ephemeral Social Media,” Sport Manage-
ment Review (in press: http://www.sciencedirect.com/science/article/pii/
S144135231730044X).
Ward, James C. and Amy L. Ostrom (2006), “Complaining to the Masses: The
Role of Protest Framing in Customer-created Complaint Web Sites,”
Journal of Consumer Research, 33, 2, 220–30.
Wilcox, Keith and Andrew T. Stephen (2012), “Are Close Friends the Enemy?
Online Social Networks, Self-esteem, and Self-control,” Journal of
Consumer Research, 40, 1, 90–103.
Zaichkowsky, Judith Lynne (1985), “Measuring the Involvement Construct,”
Journal of Consumer Research, 12, 3, 341–52.
Zhang, Kunpeng, Siddhartha Bhattacharyya, and Sudha Ram (2016), “Large-scale
Network Analysis for Online Social Brand Advertising,” MIS Quarterly, 40, 4,
849–68.