You are on page 1of 21

The current issue and full text archive of this journal is available on Emerald Insight at:

www.emeraldinsight.com/0959-6119.htm

IJCHM
31,4 Value destruction in exaggerated
online reviews
The effects of emotion, language,
1956 and trustworthiness
Melissa A. Baker
Received 28 March 2018
Revised 30 July 2018 Department of Hospitality and Tourism Management,
30 October 2018 University of Massachusetts Amherst, Amherst, Massachusetts, USA, and
6 December 2018
Accepted 20 January 2019
Kawon Kim
Department of Hotel, Restaurant and Tourism Management,
University of South Carolina, Columbia, South Carolina, USA

Abstract
Purpose – This paper aims to examine the underlying motivations, attitudes and behaviors of exaggerated
review posters and readers by examining the effect of review valence, emotional expression and language
complexity on perceived poster, website and firm trustworthiness and subsequent behavioral intentions.
Design/methodology/approach – This research uses a mixed-method approach using the qualitative
critical incident technique (CIT) and quantitative experimental design. Study 1 uses CIT to examine
exaggerated online reviews from the poster perspective where Study 2 uses CIT to examine readers’
perceptions of exaggerated reviews. Study 3 conducts a between-subjects experimental design examining the
impact of valence (positive vs negative)  emotion (low vs high)  language (vague vs detailed) on
trustworthiness and behavior intention.
Findings – Results of the two qualitative studies (Study 1 and 2) find posters and readers use language
complexity and emotions in exaggerated reviews. The results from the quantitative experimental design
study (Study 3) find that language style and emotions influence customer perceptions of poster, website and
firm trustworthiness, which also mediates the relationship between the qualitative aspects of review text on
behavioral intentions.
Practical implications – The findings provide multiple practical implications on the prevalence of
exaggerated online reviews and the importance of language and emotion in determining customer perceptions
and behavioral intentions.
Originality/value – By focusing on both readers and posters in exaggerated eWOM, specific motivations,
emotions and language, this research contributes to the literature of online reviews, customer misbehavior,
trustworthiness, language use and value co-destruction in online environments.
Keywords Language, eWOM, Value destruction, Trustworthiness, Online review,
Exaggerated review
Paper type Research paper

1. Introduction
Given the increasing competition in the hospitality and tourism industry, the investigation
International Journal of of how consumer-generated reviews affect tourism consumption decisions is incredibly
Contemporary Hospitality
Management
important. The prevalence of fake online reviews, both positive and negative, are escalating
Vol. 31 No. 4, 2019
pp. 1956-1976
and represent a growing problem theoretically and managerially (Zhang et al., 2016). Word-
© Emerald Publishing Limited of-mouth (WOM) communications are commonly assumed to be truthful, but the
0959-6119
DOI 10.1108/IJCHM-03-2018-0247 assumption is frequently wrong (Harris et al., 2016). Research throughout psychology and
sociology finds that humans regularly lie and exaggerate (Meltzer, 2003); therefore, given Exaggerated
these findings, the dearth of research into exaggerated customer WOM is perplexing (Harris online reviews
et al., 2016). Increasingly, there is evidence that customers exaggerate their electronic word-
of-mouth (eWOM) communication, but to date, empirical research that focuses on
exaggerated eWOM is limited and thus warrants further investigation (Harris et al., 2016).
Exaggerated eWOM is defined as intentionally distorted communications by customers that
misrepresent their consumption experiences. To fill these gaps in the literature, this research
examines the specific form of eWOM, exaggerated WOM in online tourism reviews. 1957
It is important for research to untangle the specific elements of review content and how it
affects customer attitudes toward the review (Folse et al., 2016). Detecting fake online
reviews is more difficult than face-to-face communication because readers can only rely on
the written words, as they cannot see the body language, facial expressions and tone, which
are critical to communication (Baker and Kim, 2018). The majority of the eWOM literature
focuses on the quantitative aspects yet the qualitative aspects, such as readability of the
text, are just as critical in affecting customer perceptions and behaviors (Agnihotri and
Bhattacharya, 2016). Customer’s perceptions of eWOM are heavily influenced by
information processing of the text language and the sentimental tone, yet this has not been
examined in the academic literature (Agnihotri and Bhattacharya, 2016; Wu et al., 2017).
Given that the content contained in online reviews contributes to attitudes and behaviors
(Folse et al., 2016), there is a need to examine the role of language in reader’s perceptions of
the content trustworthiness. As such, examining the specific language and emotionality
used are the most critical for online deceptive reviews and were thus chosen for this research
based on their importance.
Fake and exaggerated reviews undermine the credibility and trustworthiness of the
reviews as a whole (Zhang et al., 2016) yet research has not untangled which specific
constituents are affected. As the popularity of online review platforms such as Yelp and
TripAdvisor grows (Baker, 2017), so do the concerns about their credibility (Luca and
Zervas, 2016). Few studies examine the effects of eWOM on trust as perceived by the reader,
and this is surprising given that trust is a crucial driver of online interactions and should be
even more salient with the increase of deceptive reviews (Munzel, 2016). Given the rise of
documented scandals in online reviews, there needs to be a deeper investigation of consumer
perceptions of trustworthy and untrustworthy content in online reviews (Filieri, 2016).
Furthermore, as online environments involve multiple constituents such as poster, other
customers, firm and the third-party website, research is needed to extend the understanding
of how customers perceive the trustworthiness of the different constituents. The value
formation process that happens between the firm, customer and other customers can also be
destructive, which means the value is both co-created and co-destroyed (Quach and
Thaichon, 2017). Few studies examine how online reviews are a form of co-destruction and
how they affect the multiple actors involved including the poster, reader (other customer),
website and firm. Interestingly, service research finds that consumers generally trust WOM,
yet psychology research finds that most people lie at least once a day, and intentionally
exaggerate their communication (Harris et al., 2016). Given this gap in the academic
literature, it is critical to investigate trustworthiness as a key eWOM outcome.
This study aims to address the critical gaps in the literature surrounding the
examination of the co-destruction process by using a mixed-methods approach with two
qualitative studies using critical incident technique (CIT) followed by a quantitative study
using experimental design. The main motivation for adopting a mixed-method approach is
for the following reasons. First, there is a need to examine the unexplored area of
exaggerated online reviews from different stakeholders. The main challenge in empirically
IJCHM identifying review fraud is that we cannot directly observe whether a review is fake (Luca
31,4 and Zervas, 2016). That being said, it is difficult to examine the underlying motivations and
information processing procedures that both the reviewers and readers go through while
writing and reading an exaggerated review. Furthermore, there is a dearth of studies that
use qualitative methods to inductively explore consumer information processing of online
reviews, with even fewer examining trustworthiness and deception (Filieri, 2016). Therefore,
1958 it is necessary to use diverse methods to increase the underlying knowledge of exaggerated
reviews from both the poster and reader perspectives. CIT (Gremler, 2004) is an ideal
method to first investigate this phenomenon. In the first CIT, we examine the underlying
motivations, attitudes and behaviors of exaggerated review posters. The second CIT
examines the underlying motivations, attitudes and behaviors of exaggerated review
readers. Based on the results, this study proposes a conceptual model that incorporates
factors that better understand reader’s trustworthiness toward different stakeholders.
This paper is organized as follows. First, in Section 2, we present the literature pertaining
to exaggerated review and trustworthiness in an online review context to show the current
state and the gaps in the literature. Then this study presents the qualitative study results of
the two CIT studies in Section 3. Finally, building upon the existing literature combined
with the results of qualitative studies, conceptual framework and proposed hypotheses are
presented in Section 4. Discussion and conclusion are presented in Section 5.

2. Literature review
2.1 Value co-destruction
The majority of studies assume that customers act in a functional and good-mannered way
during service exchanges (Baker et al., 2012), yet the value formation process that happens
between the firm, customer and other customers can also be destructive, which means the
value is both co-created and co-destroyed (Quach and Thaichon, 2017). In online review
environments, there is a four-way interaction between the reviewer, the other customers
reading the reviews, the firm and the third party website. Where the customer misuses the
resources of the online review, it can result in co-destruction. The current literature on value
co-destruction does not provide evidence on the role and effects of multiple actors in the
online review environment. In other words, it is important to investigate how the various
actors in the service exchange such as the customer, other customer, firm and website
participate and effect value co-destruction. Increasingly, there is evidence in practice and
with key hospitality and tourism websites such as Yelp and TripAdvisor that customers
exaggerate their eWOM, but the research is extremely limited as to how these customers
represent their exaggerated eWOM in the online communication and their motives for
posting. As such, this research seeks to address this gap by examining both poster and
reader perceptions of exaggerated eWOM.

2.2 Exaggerated and fake reviews


Studies investigating deviant behavior in online contexts are limited, despite the increase in
malicious online behaviors. Fake reviews are those that are deliberately written to sound
authentic, but deceive the reader (Ott et al., 2012). Fake reviews are defined as deceptive
reviews provided with the intention to mislead the customer in their purchasing decisions
(Zhang et al., 2016). Estimations find that approximately one-third of reviews are fake or
contain falsified elements (Streitfeld, 2012). Recently, there are calls to examine specific
WOM forms, such as exaggerated WOM (Harris et al., 2016). Exaggerated WOM is defined
as intentionally distorted communications by customers to misrepresent their consumption
experiences (Harris et al., 2016) and thus constitutes a type of fake review.
Studies in sociology, psychology and anthropology find that humans habitually lie and Exaggerated
exaggerate during their interactions (Meltzer, 2003). Online platforms provide an easy online reviews
setting for creating fictitious reviews because most do not have specific restrictions on
posting and require little information (Zhang et al., 2016). In this effort, individuals who post
reviews are moving toward not simply creating entirely fake reviews based on nonexistent
experiences, but to exaggerate their actual reviews. This highlights the importance of
understanding which elements posters exaggerated in their eWOM. As previously noted,
detecting exaggerated elements is more difficult in eWOM than traditional WOM because it 1959
eliminates the body language and tone elements of communication and only relies on the
words and can therefore build upon the current research gaps.

2.3 Trustworthiness
While studies support the effect of eWOM on behavioral intentions, few studies examine the
effects of eWOM on reader perceptions of trustworthiness (Munzel, 2016). Online review
trustworthiness is particularly relevant given the global media that frequently cites
scandals, particularly related to the hospitality and tourism sector (Filieri, 2016).
Trustworthiness is the degree of confidence in the information validity in terms of
objectivity and sincerity (Hovland and Weiss, 1951). The assumption that customer WOM is
truthful is implicit in most customer WOM studies (Harris et al., 2016). However, research
finds that people regularly lie and exaggerate in their communication (Meltzer, 2003), yet
researchers find that consumers instinctively trust the comments of others (Harris et al.,
2016). Interestingly, service and marketing research is founded on the premise that
customers trust WOM, yet psychology research finds that most humans lie every day, and
intentionally exaggerate their communication (Harris et al., 2016). Given this gap in the
academic literature, it is crucial to examine trustworthiness as a key outcome of eWOM.
According to value co-destruction, the interactional process between service systems
results in a decline in at least one of the systems well-being. In other words, exaggerated
online reviews are intentionally misused which leads to value co-destruction amongst the
constituents. As up to one-third of online reviews contain fake information, exaggerated
reviews are a considerable threat to the trustworthiness of the review sites (Munzel, 2016) as
readers are increasingly concerned about untrustworthy and fraudulent information in
online reviews (Zhang et al., 2016). Exaggerated reviews undermine the credibility and
trustworthiness of the reviews as a whole (Zhang et al., 2016), yet research has not untangled
which specific targets are affected nor examined how exaggerated reviews influence the
review platform nor the target of the review. Of the limited research, the dependent variable
often focuses on perceptions of the reviewer (Kim and Gupta, 2012), and not the effects on
different constituents. This research therefore examines trustworthiness of the poster, firm,
and website as this represents a critical gap in the current literature.
Consumers are likely to be suspicious about the review poster when the text is too
extreme (Agnihotri and Bhattacharya, 2016). When consumers sense an exaggerated
component in the review, it will negatively affect trustworthiness of the poster. In addition,
online platforms and third-party review sites are an easy setting for fake and exaggerated
reviews (Zhang et al., 2016) and major sites such as TripAdvisor, Yelp, Expedia, Orbitz and
Priceline all report fake reviews as a growing problem (Ott et al., 2012; Zhang et al., 2016).
The increasing practice of falsified online content can destroy value of the review website
itself as it can decease the trustworthiness as a source of important decision making
regarding products and services. Previous research shows that people are less likely to give
attention to any media that they judge as not credible (Meltzer, 2003). This is especially
important in hospitality and tourism contexts due to the intangible nature of the sevices
IJCHM (Baker and Magnini, 2016) and that consumers or travelers can access reviews before
31,4 making decisions. Finally, the firm is often the victim of exaggerated reviews as they
contain fake information that can severely damage the firm (Zhang et al., 2016).
Increasingly, firms are outsourcing to consumers to write falsified reviews for the purpose of
either increasing the number of positive reviews of their own firm or writing negative
reviews that damage competing firms (Luca and Zervas, 2016). In fact, previous research
1960 reveals that manipulation and abuse by service firms increases consumers’ uncertainty
toward the service firm when reading reviews (Litvin et al., 2008). As such, it is important to
consider how customers perceive online reviews especially those perceived to be
exaggerated or containing fake information and how it impact their trustworthiness toward
the firm.

3. Qualitative study: critical incident technique (CIT)


CIT is a significant qualitative method that offers a rich source of data (Harrison, 2013). The
approach asks respondents to tell a story about an experience they have had (Gremler, 2004),
which serves as a basis for obtaining contextual data relevant for both management practice
as well as theory development (Breunig and Christoffersen, 2016). To explore incidents
involved in exaggerated WOM, procedures by Harris et al. (2016) were followed where
respondents were asked whether they had ever exaggerated in any of their online reviews.
With the rise on online data collection, research published in top tier journals uses critical
incidents from online written questionnaires (Baker and Kim, 2016). Accordingly, the data
were collected from Mturk which is one of the largest national online marketing panels.
According to the CIT method, it is important to let the data drive the theory (Funches, 2011).
As such, the sampling procedure continues until the data reaches a theoretical saturation
point which happens when dimensions and themes emerge, and continued data collection
does not result in new themes and adds little to conceptualized development (Baker and
Kim, 2016). Eleven incidents which did not indicate an exaggerated review in a service
industry or contained inadequate information were eliminated. The final sample contains 80
critical incidents (CIT Study 1) and 92 critical incidents (CIT Study 2).
For CIT Study 1, participants must have written and posted an exaggerated online
review. They were asked to recall and explain in detail a recent incident when they posted
an exaggerated review on social media while traveling in a tourism destination using a
structured questionnaire. We specifically asked about the incident, motivation to exaggerate
a review, emotions felt before and after posting an exaggerated review, and the expected
impact of an exaggerated review on the firm and other customers. For CIT Study 2,
participants must be consumers who have read an exaggerated review posted by another
customer. They were asked to recall and explain in detail a recent incident when they saw
what they believe was an exaggerated online review. Specific questions asked include
description of the post, expected motivation of the reviewer to exaggerate a review, emotions
in the online post, expected impact of an exaggerated review on the firm and other
customers and criteria they used to determine the trustworthiness of the review.
We followed Flanagan (1954) and other similar research in the related areas (Baker and
Kim, 2016; Funches, 2011) for data analysis. First, each response was read and reread to
familiarize the coders with the data. Second, a classification coding scheme was developed
by the researchers reading, sorting and rereading the incidents. Third, all the critical
incidents were categorized by the two independent judges. The inter-coder agreement was
92.7 per cent where inter-judge reliabilities greater than 80 per cent are considered
satisfactory. Finally, respondent’s narratives were carefully examined and grouped by the
incidents of common themes to yield the results from CIT Studies 1 and 2.
3.1 Results: CIT 1 Exaggerated
3.1.1 Type of incidents. Critical incident results showed that participants most often post an online reviews
exaggerated review when they receive poor customer service (N = 32, 34.4 per cent). Other
incidents included rude employees (21.5 per cent), long waiting time (16.1 per cent) (i.e.
exaggerating the length of waiting time), unacceptable product quality (16.2 per cent) and
unwelcoming physical environment (11.8 per cent) (i.e. dirty, unorganized servicescape).
Respondents exaggerated not only the attributes of the service failure but also other aspects 1961
that did not really happen. For example:
The waitress at an expensive restaurant had a terrible attitude, was slow and made snarky
comments to us. I was a little angry so I exaggerated how slow the service was and how bad the
food was. I said it took the waitress 30 minutes when it really only took around 15-20. The food
was pretty good but I said it was one of the worst places I have ate at.

3.1.2 Motivations for posting an exaggerated review. About half of the respondents stated
that their main motivation to write an exaggerated review was to show and release their
negative emotions (49.4 per cent):
I was extremely irritated. So, I decided to exaggerate to express my feelings so that my mind will
feel better.
In addition, respondents wanted to warn other potential customers (14.8 per cent). Twelve
per cent wanted to hurt the firm as a revenge by posting an exaggerated review:
I was angry. I wanted the restaurant to pay for making me wait. I wanted to hurt the restaurant’s
business.
Interestingly, about 10 per cent of the respondents wanted attention from other people. For
instance, respondents said that they wanted to get sympathy and agreement from others.
Also, to get more attention, some respondents made the situation sound more interesting by
exaggerating:
I was motivated to do so because I wanted the attention I would get to replies from my comments.

3.1.3 Emotions experienced. The majority of the respondents (92 per cent) stated their
emotions affected their level of exaggeration in the review. Primary emotions felt before
posting an exaggerated review included anger (58 per cent), frustration (13.5 per cent),
annoyance/irritation (12.6 per cent) and sadness (10.6 per cent):
I was irritated and annoyed with the business. I had high expectations of it when I went in and
was greatly let down.
After posting an exaggerated review, more than half of the respondents felt positive
emotions such as contentment/satisfaction (35.5 per cent), relief (27.4 per cent) and
vindication (8.4 per cent):
I was satisfied as if I have saved someone from one bad experience.
However, some respondents felt negative emotions such as annoyance (12.6 per cent), guilt
(8.4 per cent), emptiness (3.3 per cent) and regret (3.3 per cent):
I felt guilty more so than anything else. I felt like life was a joke and we all lie some way or
another.

I felt regret after posting the review. I thought maybe they were really busy that day.
IJCHM 3.1.4 Expected consequences of exaggerated review. When asked what consequences they
31,4 expected to happen to the firm after posting, 44 per cent said that they expect their
exaggerated review would negatively affect the firm’s reputation and credibility:
I felt it will surely negatively affect the service firm as I took up this matter on social platform.
In contrast, a similar percentage did not expect any impact on the firm. Some of them stated
1962 that since the firm already has high ranking and popularity, a single exaggerated review
may not affect the firm.
In terms of the expected consequences to other customers who view the exaggerated
review, 42 per cent expected that their exaggerated reviews may discourage other customers
from visiting the firm, followed by 36 per cent saying reviews may give new perspective
about the firm so that other customers may look for other reviews to make a better decision.
In other words, they believe that their review itself may not directly influence other
customers’ decision, but at least let them be aware of the firm and have a second thought
about the firm:
I think this will change other customer’s minds once they read up my post and they will choose
something else.

3.2 Results: CIT 2


3.2.1 Types of incidents. Similar to the results of CIT1, about half of the respondents
(48.5 per cent) stated that the exaggerated review read was related to poor customer
service, followed by unwelcoming physical atmosphere (22.0 per cent), unacceptable
product quality (13.8 per cent), rude employees (13.8 per cent) and long wait time (1.8
per cent).
3.2.2 Motivation. Table II shows the results of the motivation classification. Contrary to
the first CIT results, damaging the firm’s reputation was the most answered motivation
(32.6 per cent) by readers. Respondents expected that reviewers were writing an
exaggerated review as revenge toward the firm and to get people to stop going there. An
impactful finding was that 18.4 per cent thought negative exaggerated reviews were written
by paid reviewers from competitor firms. In addition, they believed some firms also
recruited individuals to write positive review about the firm:
To make the restaurant pay him to make write a good review. It is about money intention.

Possibly a competitor trying to harm the business.


Other answers showed similar results as CIT 1, which include releasing reviewer’s negative
emotions (16.3 per cent), obtaining compensation from the firm (15.3 per cent) and seeking
attention from others (7.1 per cent).
3.2.3 Emotions in exaggerated reviews. After reading an exaggerated review, more than
half of the respondents felt negative emotions toward the reviewer (52 per cent) such as
anger, annoyance and irritation due to the assumption that the reviewer is telling a lie with
malicious intent:
I just felt a little bit annoyed because he’s telling a lie.

I was irritated that they were so ignorant and ready to hurt somebody’s business.
Eighteen per cent answered that they were entertained by the exaggerated review as they
thought it was ridiculous and stupid to tell a lie and exaggerate the story:
It was kind of amusing how exaggerated it was so it kind of made me laugh. Exaggerated
Other emotions felt after reading an exaggerated review included indifference (9.0 per cent), online reviews
skepticism/suspicion (8.0 per cent), worry about if it could be true and happen to them (8.0
per cent), disappointment toward the firm for providing negative service experiences to
customers (3.0 per cent) and sympathy toward the firm for confronting such exaggerated
review (2.0 per cent).
3.2.4 Behavioral changes after reading an exaggerated review. In regards to behavioral 1963
changes, the majority of the respondents (79.5 per cent) stated that their behavior did not
change. Recognizing that it is an exaggerated review, they simply disregarded the contents
of the review. In addition, others said they became more cautious when reading reviews so
they were more likely to seek other reviews to make objective decision:
I simply ignored it which is a lie and continued reading other reviews.

No, I have seen fake reviews before: I just compared, made note, and moved on.
3.2.5 Perceptions of trustworthiness. When asked what criteria respondents use to
determine the trustworthiness of the review, 34 per cent of the respondents described that
they compare it with other reviews to decide if others had a similar situation or it’s an
outlier. Respondents said:
If the review is very different from the majority of the other reviews about the business, I tend to
be skeptical.
About 20 per cent mentioned that they use the language complexity, in which they trust a
review written with detailed information and description with specific examples of what
happened.
In a similar vein, emotional expression of the written language was another important
criterion that respondents used to determine the trustworthiness:
If it includes specific details without being overly specific and just the overall tone.

If I see something outrageous in a review, I tend to have a red flag reaction and go check out other
information to see how true it is.
Other criteria include unrealistic/extreme scenario (13.1 per cent), personal experience (10.3
per cent), reviewer’s review history (i.e. number of review) (6.5) and intuitive judgment (6.5
per cent).

3.3 Discussion of Study 1 and Study 2


The two CIT studies show consensus on the written content of exaggerated reviews. Both
exaggerated review posters and readers state that the rationale for writing an exaggerated
review was poor customer service. Approximately 50 per cent of the respondents stated the
main motivation in writing an exaggerated review was to vent negative emotions.
Interestingly, readers’ perceived motivation of exaggerated review posters was different
depending on the valence of the review. While respondents perceived the underlying
motivation of writing a negative exaggerated review was to get revenge against the firm,
positive exaggerated reviews were based on firm recruitment. In regards to specific content
criteria of deciding the trustworthiness of the review, written emotional expressions and
language complexity were found to be the most effective content. This is in line with the
results that find posters do so to vent emotions and the emotions affect their level of
IJCHM exaggeration on the review. In addition, readers perceive emotions as a negative sign of
31,4 trustworthiness of the contents of the review.
Given the qualitative results from Studies 1 and 2, there is a need to understand how
emotional expression and language each influence trustworthiness perceptions toward
different parties as well as reader’s behavioral intention. In addition, although our CIT
results show that language complexity, the length and details of the contents, is another
1964 criterion that readers use as a criterion for review trustworthiness. Language complexity is a
new construct, which has rarely been examined in the literature, yet represents a critical
construct to boundary conditions (Folse et al., 2016).

4. Study 3: experimental design


Based on the results from Studies 1 and 2, a conceptual model is presented (Figure 1) to
empirically examine the underlying motivations and written content that affect consumer’s
perception of trustworthiness in exaggerated eWOM. The two CIT studies reveal that
language complexity, which is defined as the length and details provided in the review, is
seen to be a critical indicators of trustworthiness. Second, results find that emotional
expression, defined in the marketing literature as containing exclamation marks, bold and
capitalized letters, and emojis, depicts intense emotions in written eWOM (Folse et al., 2016).
In other words, trustworthy reviews should not be overly emotional and should provide
appropriate language specificity. Although our qualitative studies affirm that review
valence, emotional expression and language complexity significantly influence review
reader’s trustworthiness toward stakeholders involved, no study yet has tested their impact
on reader’s trustworthiness followed by behavioral intentions.

4.1 Conceptual model development


4.1.1 Valence and online reviews. Online reviews are the second most trusted source of
brand information following recommendations from friends and family, yet readers are
increasingly concerned about untrustworthy and fraudulent information in online reviews
(Zhang et al., 2016). Valence describes the extent to which an experience is positive or
negative and is one of the most imperative factors in online reviews (Baker, 2017). Studies
typically focus on one valence, but examining negative and positive valence content is
needed (Agnihotri and Bhattacharya, 2016; Folse et al., 2016). Of the limited studies, most
examine negative review content. Furthermore, literature on valence effects finds that
negatively valenced reviews have an adverse effect on reviewers in terms of helpfulness and
trustworthiness of the reviewed object (Folse et al., 2016), and negative reviews are more
influential to consumer behavior than positive reviews. However, the content contained in

Review
Valence Trustworthiness
toward poster

Behavioral intention
toward website
Emotional Trustworthiness
expression toward website

Behavioral intention
toward firm
Trustworthiness
Figure 1. Language toward firm
Conceptual model complexity
both positive and negative reviews can significantly impact reader perceptions. Therefore, Exaggerated
this study builds upon the research in marketing journals not only to examine the effect of online reviews
language on negative reviews (Folse et al., 2016) but also to find how language complexity
and emotions affect negative and positively valence reviews differently. Therefore, this
research hypothesizes:

H1. Review valence has a positive effect on (a) trustworthiness toward the poster, (b)
trustworthiness toward the website and (c) trustworthiness toward the firm. 1965
4.1.2 Emotions and online reviews. Dysfunctional customers act in a way that causes
problems for the firm, employees and other customers and their behavior is often linked to
emotions or personality characteristics (Baker et al., 2012). Human interaction studies find
that anger and frustration are linked to dishonest communication and can be used to justify
falsified communication (Meltzer, 2003). Examining emotions as a distinct dimension of
exaggerated eWOM is important and can build upon academic research, as less is known
about how emotions play a role in exaggerated or falsified reviews (Luca and Zervas, 2016).
Misbehaving customers do so (in part) to share their emotions of anger, rage, and
frustration, which can affect other customers in the service environment and venting these
emotions can affect the various constituents.
A review can have similar valence, but possess different emotionality (Ren and Nickerson,
2014). As such, it is important to consider valence and emotionality separately. According to
conventions of expressing written emotions (Kim and Gupta, 2012) emotions are added using
capital letters, explanation points, and a phrase describing the reviewer’s internal emotional state.
Utility of reviews may decline because consumers might discount information at an extremely
emotional level (Agnihotri and Bhattacharya, 2016). Stated differently, with exaggerated eWOM,
consumers realize that extremes in emotions signal doubt (Liljander et al., 2015). Emotional
expressions are expected to influence trustworthiness (Folse et al., 2016). A review that is too
emotionally expressive would be cognitively processed as an attempt to manipulate the
consumers purchasing decisions (Agnihotri and Bhattacharya, 2016). More specifically, the use of
emotions in written online reviews will adversely impact perceptions of trustworthiness:

H2. Emotional expression has a negative effect on (a) trustworthiness toward the poster,
(b) trustworthiness toward the website and (c) trustworthiness toward the firm.
4.1.3 Language and online reviews. One of the most important cues that respondents use to
assess review trustworthiness is the language. The language used in reviews is critical in
determining fake versus authentic reviews (Zhang et al., 2016). One approach in
distinguishing between authentic and fictitious reviews is language-based detection
(Banerjee et al., 2017). Surprisingly, the impact of the language used, and the impact of the
expressions is largely unaddressed (Folse et al., 2016), yet language may be critical in
shaping customer evaluations (Wu et al., 2017).
This is grounded in information manipulation theory which expects reviews to differ in
clarity and specificity. Information manipulation theory also suggests authentic reviews are
more specific than fictitious ones where specificity refers to the extent to which the reviews
are detailed (Banerjee et al., 2017). Reviews perceived as untrustworthy are very short and
do not provide specific details of the reviewer’s experience. More specifically, the use of
descriptive information versus vague information separates trustworthy from suspicious
reviews (Munzel, 2016). For instance, Salehan and Kim (2016) show that the length of a
review positively affects reader’s helpfulness. Extending this logic, we hypothesize:
IJCHM H3. Language complexity has a positive effect on (a) trustworthiness toward the
31,4 poster, (b) trustworthiness toward the website and (c) trustworthiness toward the
firm.
4.1.4 Trustworthiness and behavioral intentions. As noted by communication experts,
emotional language is an important type of message element that inhibits or promotes
persuasion (Folse et al., 2016). By examining the words used in online reviews, we can better
1966 understand how online reviews affect customer behavior (Ren and Nickerson, 2014).
Trustworthiness in marketing is found to influence attitude and subsequent behavioral
intentions (Ayeh et al., 2013). Likewise, perceptions of trustworthiness in online contexts are
likely to influence customer behavioral intentions. Review comprehensibility and
sentimental content can lead other customers to perceive those reviews as untrustworthy
(Agnihotri and Bhattacharya, 2016). In terms of value co-destruction, it is important to
examine how perceptions of trustworthiness of different constituents affect behavioral
intentions toward the firm and website. Competition is rising in terms of sources of
information in online contexts. It is therefore critical for websites to consider how customers
perceive their trustworthiness when the read fake reviews on their site. In addition, firms
must be cognizant of how the trustworthiness of different constituents affects customer
behavioral intentions. Therefore, this research proposes:

H4. Trustworthiness toward the poster has a positive effect on (a) behavioral intention
toward firm and (b) behavioral intention toward website.
H5. Trustworthiness toward the website has a positive effect on (a) behavioral intention
toward firm and (b) behavioral intention toward website.
H6. Trustworthiness toward the firm has a positive effect on (a) behavioral intention
toward firm and (b) behavioral intention toward website.

4.2 Research design and sample


A 2 (Valence: positive vs negative)  2 (Emotion: high vs low)  2 (Language: vague vs
detail) between-subjects scenario-based experimental design is used. We use
TripAdvisor because it is one of the largest online review and travel-related user-
generated website that contains nearly 600 million reviews. Second, fake reviews are a
growing problem of mainstream review sites such as Yelp and TripAdvisor. TripAdvisor
does not use a filtering algorithm for falsified reviews (Zhang et al., 2016), which make it
an ideal context, with high practical implications as the context for the study. In addition,
due to its popularity, it enhances the external validity of the research method, better
reflects the reality and enhances the likelihood of showing real behavior of the
respondents.
A total of 341 participants were recruited from Amazon Mechanical Turk, which features
a diverse nation-wide pool of consumers and provides high quality data (Buhrmester et al.,
2011). Hair et al. (1998) regards five respondents per variable to be analyzed as the lower
limit, but the most acceptable way of determination is 10:1 ratio (10 samples for one
variable). As Study 3 includes eight variables, 341 participants is an adequate sample size to
meet the criteria. As this research examines an online review scenario, we asked
respondents the frequency of reading online reviews as a screening question and eliminated
participants who answered rarely. About 60 per cent of respondents frequently read online
reviews. The average age of a respondent was 33 years with the range of 20 to 72 years, and
44 per cent were females. More than half of the respondents were white (59.6 per cent), and Exaggerated
about 50 per cent had college degree. online reviews

4.3 Experimental stimuli and procedures


The stimuli reflected a hypothetical scenario in which participants were asked to imagine a
situation while searching for hotel information in an online travel review website for an
1967
upcoming trip, the participant noticed an online review. After showing a hypothetical online
review, participants are asked to fill out a survey questionnaire by referring only to the
scenario provided. Participants were randomly assigned to one of the eight conditions where
they read an online review with different level of valence, emotion and language complexity.
Review valence was manipulated by either a negative or positive description of the
reviewers as well as star rating adopting from Ren and Nickerson (2014). Positive words
such as best, great, excellent were used to describe the positive attributes of the hotel while
negative words such as worst, bad, terrible were used to describe the negative attributes of
the hotel. In addition, we used onestar rating to represent negative valence and a five-star
rating to represent positive valence. Following Kim and Gupta (2012), emotion was
manipulated by emotional words (i.e. hate, angry, love, happy) describing the reviewer’s
internal emotional states, capitalize letters and exclamation marks in the review.
Specifically, the high emotion review read, “BEST hotel ever!!!!!!!: Everything about the
experience was EXCELLENT!!!!!!! I really LOVE this hotel!!!!!!! This experience made me
SOOOOOOO HAPPY!!!!!!!” and “Bad hotel experience: Everything about the experience was
terrible. Overall, the hotel was not very good” was the low emotion negative review. Finally,
the manipulation for language complexity was accomplished through describing the degree
of detail in the review. In doing so, we referred the attributes that TripAdvisor asks
reviewers to rate in addition to general written review including service, location and room.
Specifically, in the detailed language condition, the review read as followed:
Everything about the experience was excellent. We enjoyed the hotel’s fabulous location which
was convenient to shopping, sites and restaurants. Our room was spacious, comfortable and
clean. The service was exceptional too. Sam, part of the concierge team, greeted us every evening
by name. He went out of his way to make us feel at home. Overall, the hotel is very good.
In the vague language condition, the review stated the following: “Everything about the
experience was excellent. Overall, the hotel is very good”.

4.4 Measures
All items were measured using a seven-point Likert-type scale. As a manipulation check,
review valence was measured by four items (Kim and Gupta, 2012). Participants perception
of emotion on the review was measured by three items (Kim and Gupta, 2012; Ren and
Nickerson, 2014) (a = 0.89). Language complexity was adapted from Banerjee et al., 2017
using four items (a = 0.88). To measure trustworthiness toward the poster (a = 0.95), the
review website (a = 0.96) and the firm (a = 0.93), three items from Andrews et al. (1998) were
used for each. Behavioral intention toward the firm was measured with two items (a = 0.95).
Behavioral intention toward the review website was measured with two items adopted from
Ayeh et al. (2013) (a = 0.93). Participants online review skepticism was measured with a
four-item (a = 0.89). Factor analysis and reliability tests with full measurement items are
presented in Table I.
IJCHM Constructs and items Loading a
31,4
Valence (1: Strongly disagree; 7: Strongly agree) 0.98
The review was negative 0.97
The review was positive 0.97
The review received a negative rating 0.97
The review received a positive rating 0.98
1968
Emotion (1: Strongly disagree; 7: Strongly agree) 0.89
The review demonstrated a high level of emotion 0.94
The review demonstrated a low level of emotion 0.86
The emotion level used in the review was (1: very low level emotion - 7: 0.92
very high level emotion)
Language complexity (1: Strongly disagree; 7: Strongly agree) 0.88
The language used was detailed 0.88
The language used was vague 0.83
The length of the review was long enough to explain what really happened 0.85
The length of the review was too short to explain what really happened 0.85
Review skepticism (1: Strongly disagree; 7: Strongly agree) 0.89
I am basically doubtful about online reviews 0.82
Online reviews are often questionable 0.83
I am generally uncertain about online reviews 0.85
I am generally skeptical about online reviews 0.89
Trustworthiness toward the poster (1: Strongly disagree; 7: Strongly agree) 0.95
The online review from the poster is trustworthy 0.93
The online review from the poster is credible 0.92
The online review from the poster is believable 0.89
Trustworthiness toward the review website (1: Strongly disagree; 7: 0.96
Strongly agree)
Based on this review, the online review from the review website is 0.95
trustworthy
Based on this review, the online review from the review website is credible 0.92
Based on this review, the online review from the review website is 0.91
believable
Trustworthiness toward the firm (1: Strongly disagree; 7: Strongly agree) 0.93
Based on this review, the hotel described in the review is trustworthy 0.86
Based on this review, the hotel described in the review is credible 0.93
Based on this review, the hotel described in the review is believable 0.84
Behavioral intention toward the firm 0.95
After reading the online review, how likely are you to choose this hotel? (1: 0.97
Very unlikely; 7: Very likely)
After reading this online review, how interested are you to visit at this 0.93
hotel? (1: Not at all; 7: Extremely)
Behavioral intention toward review website (1: Not at all; 7: Extremely) 0.93
Table I. After reading this online review, how interested are you to visit at this 0.84
Constructs and items review website in the future to seek travel advice?
with factor loadings After reading this online review, how interested are you to use the content 0.96
and reliabilities at this review website in the future?
4.5 Results Exaggerated
4.5.1 Manipulation checks. A series of independent sample t-tests were used to check online reviews
whether respondents perceived the stimuli as intended. Participants rated the positive
review (M = 6.37, SD = 1.23) and negative review (M = 1.57, SD = 1.04), indicating
successful valence manipulation (t = 38.94, p < 0.001). The high emotion review (M = 6.12,
SD = 1.07) and low emotion review (M = 4.36, SD = 1.44) (t = 12.65, p < 0.001) were also
significant. Language complexity was also significant indicating differences
between detailed (M = 4.93, SD = 1.34) and vague language (M = 2.70, SD = 1.42) used in the
1969
review (t = 14.76, p < 0.001). Taken together, these results indicate that the manipulation of
each of the three manipulated cues was successful.
4.5.2 Trustworthiness. A three-way MANCOVA tested the effects of three review
attributes on perceived trustworthiness toward the poster, review website and the firm by
controlling frequency of usage of review website and review skepticism (Table II). The
results show that valence, emotion and language complexity have significant effects
on trustworthiness [F(3,326) = 15.86, p < 0.001, h 2 = 0.87; F(3,326) = 5.42, p < 0.01, h 2 = 0.95;
F(3,326) = 4.71, p < 0.01, h 2 = 0.96]. Univariate results reveal significant effects of valence on
trustworthiness toward the firm, F(1, 328) = 40.13, p < 0.001, supporting H1(a). Participants
are more likely to trust the firm if it is positively valenced (M = 4.950, SE = 0.09) than
negatively valenced (M = 4.10, SE = 0.10). However, no significant effect
exists on trustworthiness toward the poster and the firm, failing to support H1(b) and H1(c).
In addition, there was a significant effect of emotion on trustworthiness toward the poster
[F(1,328) = 11.10, p < 0.01], the website [F(1,328) = 4.61, p < 0.05], and the firm
[F(1,328) = 7.91, p < 0.01], supporting H2(a), H2(b) and H2(c). Participants have a high

Multivariate Univariate
Source Wilks’ lambda p-value DV F p-value

Valence 0.873 0.000 TTP 2.893 0.090


TTW 2.529 0.113
TTF 40.133 0.000
Emotion 0.953 0.001 TTP 11.097 0.001
TTW 4.610 0.033
TTF 7.910 0.005
Language 0.958 0.003 TTP 10.081 0.002
TTW 12.482 0.000
TTF 9.533 0.002
Valence  Emotion 0.173 0.914 TTP 0.017 0.897
TTW 0.181 0.671
TTF 0.011 0.917
Valence  Language 1.938 0.123 TTP 1.849 0.175
TTW 3.900 0.049
TTF 3.052 0.082
Emotion  Language 1.958 0.116 TTP 3.793 0.052
TTW 2.310 0.129
TTF 0.004 0.950
Valence  Emotion  Language 4.603 0.004 TTP 0.019 0.891
TTW 0.002 0.962
Table II.
TTF 9.027 0.003
MANCOVA and
Notes: TTP = Trustworthiness toward poster; TTW = Trustworthiness toward website; TTF = univariate follow-up
Trustworthiness toward firm results for Study 3
IJCHM trustworthiness toward the poster when the review shows low level of emotions (M = 4.80,
31,4 SE = 0.10) than high level of emotions (M = 4.30, SE = 0.11). Similar patterns were shown in
both trustworthiness toward the review website, increase of trust in low level of emotion
(M = 4.80, SE = 0.10) than in high level of emotion (M = 4.48, SE = 0.11), and trustworthiness
toward the firm, higher level of trustworthiness in low level of emotion (M = 4.71, SE = 0.09),
compared to high level of emotion (M = 4.34, SE = 0.10).
1970 Similarly, language had a significant effect on trustworthiness toward the poster
[F(1,328) = 10.08, p < 0.01], the website [F(1,328) = 12.48, p < 0.001] and the firm [F(1,328) =
9.533, p < 0.01], supporting H3(a), H3(b) and H3(c). Participants have a high
trustworthiness toward the poster when the review includes detailed information (M = 4.79,
SE = 0.11) than vague information (M = 4.30, SE = 0.11). Similar patterns were shown in
both trustworthiness toward the review website, increase of trust in detailed information
condition (M = 4.90, SE = 0.11) than vague information condition (M = 4.38, SE = 0.10), and
trustworthiness toward the firm, higher level of trustworthiness in detailed information
condition (M = 4.73, SE = 0.10), compared to vague information condition (M = 4.32, SE =
0.09).
In terms of the interaction effect, valence interacted with language to influence
trustworthiness toward the website [F(1,328) = 3.90, p < 0.05]. As indicated in Figure 2,
when the review is stated with detailed information, participants had greater trust toward
the review website under negative valence condition (M = 4.93, SE = 0.15) compared to
positive valence condition (M = 4.88, SE = 0.16). In contrast, when the review is stated with
vague information, the reverse patterns were shown. Participants were more likely to trust
the review website under positive valence condition (M = 4.64, SE = 0.13) compared to
negative valence condition (M = 4.11, SE = 0.16). In addition, even though the effect was
marginally significant, similar effect was shown on trustworthiness toward the firm. That
being said, although providing detailed information is worthy for increasing reader’s
trustworthiness toward the website and the firm, it might not be the case for positive review
as people tend to have skepticism about the review detailing out too much on positive
things.
Interestingly, there was a significant three-way interaction among valence, emotion, and
language on trustworthiness toward the firm [F(1,328) = 9.03, p < 0.01]. As shown in
Figure 3, under negative valence condition, the difference in trustworthiness toward the firm
as a result of language complexity was larger under high emotion condition (Mvauge = 3.40
vs Mdetail = 4.44) than low emotion condition (Mvauge = 4.15 vs Mdetail = 4.41). Under the
positive valence condition, inclusion of detail information helped to increase trustworthiness
toward the firm with low emotion (M = 5.44), while not in high emotion condition (M = 4.64).
On the other hand, inclusion of vague information did not show any differences between low

5
Vague
Trustworthiness toward

4.8 Detailed

4.6
website

4.4

4.2
Figure 2.
4
Interaction effect of
valence and language 3.8
Negative Positive
on TTW
Valence
Negative valence Exaggerated

Trustworthiness toward firm


4.6
Vague online reviews
4.4 Detailed
4.2
4
3.8
3.6
3.4
3.2
3
1971
Low High
Emotion

Positive valence
Trustworthiness toward firm

5.6
Vague
5.4
Detailed
5.2
5
4.8
4.6 Figure 3.
4.4 Three-way
4.2 interaction effect
Low High
on TTF
Emotion

emotion condition (M = 4.85) and high emotion condition (M = 4.86). To be more specific,
when it is a positive review, while providing detailed description information of the incident
is helpful to enhance trustworthiness toward firm when combined with low emotion, it can
backfire when combined with high emotion. On the other hand, providing detailed
description of the incident is always helpful to increase trustworthiness toward firm
regardless of the level of emotion for a negative review. In terms of the covariates, both
frequency of usage of review website [F(3,326) = 5.16, p < 0.01] and review skepticism
[F(3,326) = 8.06, p < 0.001] reached the level of statistical significance. Future research is
needed to examine the impact frequency of usage of review website and review skepticism
on reader’s perceived trustworthiness.
In addition, given that our qualitative research results support that too extreme review
valence accompanying with tensed emotional expression affects perceptions of
trustworthiness (Filieri, 2016), as an additional analysis, by using manipulation check
questions as a latent variable (classified group into three group, negative, neutral and
positive valence), we tested whether an inverted U-shape relationship exists between
valence  emotional expression on trustworthiness. Results support a U-shape relationship
under high emotional expression, while not under low emotional expression. That being
said, in the high emotional expression condition, perceived trustworthiness toward poster,
website and firm was lower when participants perceive review as either extremely negative
or positive than neutral (Figure 4).
4.5.3 Behavioral intentions. To investigate the effect of reader’s perceived
trustworthiness toward the poster, website and the firm on behavior intention, a series of
multiple regression analyses were conducted. The results show that firm trustworthiness
has a significant impact on firm behavioral intention ( b = 0.725, p < 0.001), supporting H6a,
while poster and website trustworthiness did not affect behavioral intention toward the
firm, failing to support H4a and H5a. Second, behavioral intention toward the website was
regressed on poster, website and firm trustworthiness. The results show that
IJCHM trustworthiness toward the website has a significant impact on behavioral intention toward
31,4 the website ( b = 0.514, p < 0.001), supporting H5b, while trustworthiness toward the poster
and the firm did not affect behavioral intention toward the website, failing to support H4b
and H6b.
4.5.4 Mediation effect. To further examine the underlying mechanism, indirect effects
were tested with the significant path. Following Preacher and Hayes (2008), results from
1972 bootstrapping procedure that generated a sample size of 5,000 reveal a mediation effect of
firm trustworthiness between valence (95 per cent CI = [0.1833, 0.4991]), between emotion
(95 per cent CI = [0.3599, 0.0207]), and between language (95 per cent CI = [0.0216,
0.3674]) and behavioral intentions toward the firm Similarly, website trustworthiness also
mediated the relationship between emotion and behavioral intention toward the website (95
per cent CI = [0.3791, 0.0153]) and between language and behavioral intention toward
the website (95 per cent CI = [0.1017, 0.4585]).

5. Discussion and conclusion


5.1 Theoretical implications
This research contributes to theory building in several ways. First, while research is
examining the impact of co-creation, very little research examines the effect of
co-destruction in service environments. However, this is critical to investigate as value is
formed or destroyed by the actions of the various actors. The majority of the eWOM
research only investigates the impact on the reader or the firm. Very few studies, if any,
examine the co-destructive effect on all the constituents involved in eWOM. In online
environments, this refers to the customers who post, other customers who read and view the
website, and the firm. As such, this study first contributes to the literature by examining
this underdeveloped yet critical aspect of co-destruction in online environments.
Second, the results contribute to the vital literature on the quality of user-generated
content (Luca and Zervas, 2016) especially, as it relates to exaggerated reviews. While the
majority of the literature on review fraud comes from computer science and focuses on data-
mining algorithms, this research focuses on a management and marketing perspective,
noting the effects from a value destruction standpoint. There is a prevalence of customers
exaggerating their consumption experiences, yet empirical research that focuses on this
phenomenon is lacking (Harris et al., 2016). This research contributes to the area of
exaggerated reviews through two qualitative CIT studies to provide depth of content from
both the poster and reader perspectives. The first focuses on the scope of customers that

5.2

4.8
Trustworthiness

4.6

4.4 Low emotion


High emotion
4.2
Figure 4.
Inverted U-shape 4
relationship between
valence x emotional 3.8
Negative Neutral Positive
expression
Valence
engage in exaggerated eWOM and the specific motivations and elements used within these Exaggerated
reviews. The second examines other customer perceptions who read these exaggerated online reviews
reviews. This builds upon theory by being one of the first studies to more deeply examine
exaggerated online reviews from a management perspective.
Third, the majority of the online review literature focuses on the quantitative characteristics,
while ignoring the qualitative characteristics of review content, especially language and
emotions which are just as critical in affecting customer perceptions and behaviors (Agnihotri
and Bhattacharya, 2016). Therefore, this study fills this gap by investigating the relationship 1973
between online exaggerated reviews and the language and emotions used. More specifically,
CIT 1 finds that posters of exaggerated reviews do so in part to vent emotions. CIT finds that
readers of exaggerated fake posts find the utility and trustworthiness of the reviews
diminished as a result of exaggerated emotions and language present.
Fourth, as the use and popularity of online review platforms increases (Baker, 2017), so
do concerns regarding the trustworthiness related to exaggerated reviews. The study
supports that emotions and language complexity play an important role in forming reader’s
perceptions of trustworthiness of poster, website and firm. Interestingly, although providing
detailed information is worthy for increasing reader’s trustworthiness toward the website
and the firm, it might not be the case for positive review as people tend to be suspicious
about the review detailing out too many positive things. Furthermore, perceptions of
website trustworthiness significantly affect behaviors toward the website and
trustworthiness toward the firm significantly affect behavioral intention toward the firm. In
negatively valenced reviews, when the language complexity is detailed, there is no
significant difference between high and low emotion in the written review on firm
trustworthiness. However, when there is vague language complexity, there is a significant
difference between the high and low emotions. Specifically, there is the lowest
trustworthiness in negative reviews when the language is vague and contains high emotion.
Conversely, for positively valenced reviews, there is no difference in trustworthiness to the
firm when there is vague language complexity. In addition, firm trustworthiness is the
lowest when there is detail and high emotion. This shows an opposite pattern for positive
and negatively valenced reviews. In other words, language complexity yields higher
trustworthiness than vague language, but only for negatively valenced reviews.
Fifth, there is a call for more research that uses mixed methods, as they increase the
internal and external validity of studies (Fong et al., 2016; Harrison, 2013). Through the use
of experiments, insights are gained into the elements to detect exaggerated online reviews.
In other words, experimental design allows for the examination of causal impacts where the
key variables can be manipulated experimentally. While the experimental design provides a
high amount of internal validity, the CIT examination of both exaggerated posters and other
customer who read the exaggerated posts increases the methodological significance of this
research by balancing high internal validity (experimental design) and examination of the
real world marketplace (consumer posts) to increase external validity. As such, this research
contributes methodologically and theoretically by incorporating mixed qualitative and
quantitative methods which are underused in management research.

5.2 Practical implications


In accordance with previous literature, our study also reveals that positive comments about
a firms’ products and services can bring positive attitudinal changes in consumers’
perceptions and negative feedback can bring about negative attitudinal and behavioral
changes (Agnihotri and Bhattacharya, 2016). Given that the valence of online reviews plays
an important role, firms should carefully monitor the valence of reviews, especially for the
IJCHM negative valence reviews. However, the results also find that the qualitative aspects of the
31,4 text, namely language style and emotions, can influence customer perceptions of
trustworthiness and subsequent attitudes toward to the firm and website. Given these
results, it is vital for marketers and management to understand the significance of text
analytics. This is because text analytics, such as review content characteristics, can enable
marketers to quantify unstructured data (Agnihotri and Bhattacharya, 2016).
1974 The findings of this research provide multiple practical implications regarding the
importance of exaggerated eWOM and the criticality of creating trustworthy online review
platforms. Using machine learning software and platforms can assist in detecting fake
online reviews using the language provided (Zhang et al., 2016). In other words, firms and
websites should seek to develop and use computer science tools and algorithms to identify
fraudulent online reviews. While a majority of studies from a computer science perspective
use star rating and user characteristics in determining fraudulent reviews, the results of this
study contribute to managerial practice by finding that qualitative elements such as
language and emotions in the text should also be considered in the algorithms.
As language style and emotions significantly impact perceptions of trustworthiness, this
research recommends website and forum managers provide relevant guidelines for online
review writing. In addition, some tourism and hospitality companies selectively highlight
reviews based on advice that emotional reviews are more eye catching (Wu et al., 2017). The
findings of this research suggest that overly emotional reviews are found to be less
trustworthy and thus should be highlighted with caution. Although many factors related to
exaggerated eWOM are unknown and out of a firms control, there are factors within their
control. This includes communicating with customers within the servicescape to minimize
venting and dissatisfaction. By providing opportunities within the servicescape for venting,
this can reduce exaggerated negative WOM (Harris et al., 2016).

5.3 Limitations and future research


The findings of our research are subject to some limitations and provide suggestions for future
research. First, it is very difficult to ascertain whether reviews are truly fake. As such,
this research focused on exaggerated reviews, which contain fake and falsified elements.
Future research should continue refine deviant online behavior and examine differences
resulting from different types of fake, falsified, and exaggerated eWOM content. Second, this
research examined co-destruction in an online setting with hotel reviews, which may not have
generalizability across different social media, online and in-person contexts. Therefore, more
research is needed that examines the effect of co-destruction across different online and in
person contexts as well as across different services such as retail, restaurant, hotel, and tourism.
Third, based on the results of the CIT analyses, this research empirically examined two aspects
of emotional and language. Noting that written communication is complex, future research
should seek to examine additional language, emotion, and written elements in eWOM as well
as in person feedback. This is particularly important as written language lacks the tone and
body language of spoken language, thus disentangling the elements of the emotions embedded
in online reviews is even more critical. Fourth, although our study used TripAdvisor as a study
context to test our conceptual model, we suggest future research to examine the impact on
website familiarity and its impacts on trustworthiness. Although this study focuses on
examining four stakeholders such as poster, reader, online review website and firm, future
research can incorporate other parties that could be influenced by an exaggerated online review
(i.e. transactional website; official hotel website, DMOs). Finally, future research should further
expand on the results to find potential differences between poster, website and firm
trustworthiness as boundary conditions.
References Exaggerated
Agnihotri, A. and Bhattacharya, S. (2016), “Online review helpfulness: role of qualitative factors”, online reviews
Psychology and Marketing, Vol. 33 No. 11, pp. 1006-1017.
Andrews, J.C., Netemeyer, R.G. and Burton, S. (1998), “Consumer generalization of nutrient content
claims in advertising”, The Journal of Marketing, Vol. 62 No. 4, pp. 62-75.
Ayeh, J.K., Au, N. and Law, R. (2013), “Do we believe in TripAdvisor? Examining credibility
perceptions and online travelers’ attitude toward using user-generated content”, Journal of
Travel Research, Vol. 52 No. 4, pp. 437-452.
1975
Baker, M.A. (2017), “Electronic customer relationship management and customer satisfaction”, The
Routledge Handbook of Consumer Behaviour in Hospitality and Tourism, Routledge, Abingdon,
pp. 396-403.
Baker, M.A. and Kim, K. (2016), “Other customer service failures: emotions, impacts, and attributions”,
Journal of Hospitality and Tourism Research, Vol. 42 No. 7, pp. 1067-1085.
Baker, M.A. and Kim, K. (2018), “The role of language, appearance, and smile on perceptions of
authenticity versus rapport”, International Journal of Hospitality Management, Vol. 74, pp. 171-179.
Baker, M.A. and Magnini, V.P. (2016), “The evolution of services marketing, hospitality marketing and
building the constituency model for hospitality marketing”, International Journal of
Contemporary Hospitality Management, Vol. 28 No. 8, pp. 1510-1534.
Baker, M.A., Magnini, V.P. and Perdue, R.R. (2012), “Opportunistic customer complaining: causes,
consequences, and managerial alternatives”, International Journal of Hospitality Management,
Vol. 31 No. 1, pp. 295-303.
Banerjee, S., Chua, A.Y. and Kim, J.J. (2017), “Don’t be deceived: using linguistic analysis to learn how to
discern online review authenticity”, Journal of the Association for Information Science and
Technology, Vol. 68 No. 6, pp. 1525-1538.
Breunig, K.J. and Christoffersen, L. (2016), “If x then why? Comparative analysis using critical incidents
technique”, Journal of Business Research, Vol. 69 No. 11, pp. 5141-5146.
Buhrmester, M., Kwang, T. and Gosling, S.D. (2011), “Amazon’s mechanical turk: a new source of
inexpensive, yet high-quality, data?”, Perspectives on Psychological Science: A Journal of the
Association for Psychological Science, Vol. 6 No. 1, pp. 3-5.
Filieri, R. (2016), “What makes an online consumer review trustworthy?”, Annals of Tourism Research,
Vol. 58, pp. 46-64.
Flanagan, J.C. (1954), “The critical incident technique”, Psychological Bulletin, Vol. 51 No. 4, pp. 327-358.
Folse, J.A.G., Porter, M., III, Godbole, M.B. and Reynolds, K.E. (2016), “The effects of negatively
valenced emotional expressions in online reviews on the reviewer, the review, and the product”,
Psychology and Marketing, Vol. 33 No. 9, pp. 747-760.
Fong, L.H.N., Law, R., Tang, C.M.F. and Yap, M.H.T. (2016), “Experimental research in hospitality and
tourism: a critical review”, International Journal of Contemporary Hospitality Management,
Vol. 28 No. 2, pp. 246-266.
Funches, V. (2011), “The consumer anger phenomena: causes and consequences”, Journal of Services
Marketing, Vol. 25 No. 6, pp. 420-428.
Gremler, D.D. (2004), “The critical incident technique in service research”, Journal of Service Research,
Vol. 7 No. 1, pp. 65-89.
Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E. and Tatham, R.L. (1998), Multivariate Data Analysis,
Prentice Hall, Upper Saddle River, NJ.
Harris, L.C., Fisk, R.P. and Sysalova, H. (2016), “Exposing Pinocchio customers: investigating
exaggerated service stories”, Journal of Service Management, Vol. 27 No. 2, pp. 63-90.
Harrison, R.L. (2013), “Using mixed methods designs in the journal of business research, 1990-2010”,
Journal of Business Research, Vol. 66 No. 11, pp. 2153-2162.
IJCHM Hovland, C.I. and Weiss, W. (1951), “The influence of source credibility on communication
effectiveness”, Public Opinion Quarterly, Vol. 15 No. 4, pp. 635-650.
31,4
Kim, J. and Gupta, P. (2012), “Emotional expressions in online user reviews: how they influence
consumers’ product evaluations”, Journal of Business Research, Vol. 65 No. 7, pp. 985-992.
Liljander, V., Gummerus, J. and Soderlund, M. (2015), “Young consumers’ responses to suspected covert
and overt blog marketing”, Internet Research, Vol. 25 No. 4, pp. 610-632.
1976 Litvin, S.W., Goldsmith, R.E. and Pan, B. (2008), “Electronic word-of-mouth in hospitality and tourism
management”, Tourism Management, Vol. 29 No. 3, pp. 458-468.
Luca, M. and Zervas, G. (2016), “Fake it till you make it: reputation, competition, and yelp review
fraud”, Management Science, Vol. 62 No. 12, pp. 3412-3427.
Meltzer, B. (2003), “Lying: deception in human affairs”, International Journal of Sociology and Social
Policy, Vol. 23 Nos 6/7, pp. 61-79.
Munzel, A. (2016), “Assisting consumers in detecting fake reviews: the role of identity information
disclosure and consensus”, Journal of Retailing and Consumer Services, Vol. 32, pp. 96-108.
Ott, M., Cardie, C. and Hancock, J. (2012), “Estimating the prevalence of deception in online review
communities”, Proceedings of WWW 2012, pp. 201-210.
Preacher, K.J. and Hayes, A.F. (2008), “Asymptotic and resampling strategies for assessing and
comparing indirect effects in multiple mediator models”, Behavior Research Methods, Vol. 40
No. 3, pp. 879-891.
Quach, S. and Thaichon, P. (2017), “From connoisseur luxury to mass luxury: value co-creation and co-
destruction in the online environment”, Journal of Business Research, Vol. 81, pp. 163-172.
Ren, J. and Nickerson, J.V. (2014), “Online review systems: how emotional language drives sales”,
Twentieth Americas Conference on Information Systems, available at: https://ssrn.com/
abstract=2426694
Salehan, M. and Kim, D.J. (2016), “Predicting the performance of online consumer reviews: a sentiment
mining approach to big data analytics”, Decision Support Systems, Vol. 81, pp. 30-40.
Streitfeld, D. (2012), “For $2 a star, an online retailer gets 5-star product reviews”, available at: www.
nytimes.com/2012/01/27/technology/for-2-a-star-aretailer-gets-5-star-reviews.html
Wu, L., Shen, H., Fan, A. and Mattila, A.S. (2017), “The impact of language style on consumers 0
reactions to online reviews”, Tourism Management, Vol. 59, pp. 590-596.
Zhang, D., Zhou, L., Kehoe, J.L. and Kilic, I.Y. (2016), “What online reviewer behaviors really matter?
Effects of verbal and nonverbal behaviors on detection of fake online reviews”, Journal of
Management Information Systems, Vol. 33 No. 2, pp. 456-481.

Further reading
Kim, K. and Baker, M.A. (2017), “How the employee looks and looks at you: building customer-
employee rapport”, Journal of Hospitality and Tourism Research, Vol. 43 No. 1, pp. 20-40.

Corresponding author
Melissa A. Baker can be contacted at: mbaker@isenberg.umass.edu

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com

You might also like