Buying opinions?
The impact of rewards and platform type on the composition of online
reviews.
Master’s thesis T.C.D. Schilder Student number 10245278
Graduate School of Communication Master’s programme Communication Science
Track Persuasive Communication Dr. L. M. Willemsen
Abstract
Marketers increasingly offer rewards in return for an online review, which has led to a body of research into outcome variables such as sales and awareness. This study takes a writers’ perspective and investigates the difference in composition of online reviews for rewarded and unrewarded consumers, taking into account the moderating effect of platform type on which the review is posted. Based on previous research it is hypothesized that rewarding leads to a composition with more extreme ratings, more negative valence, less rational arguments, more punctuation cues and shorter reviews. This effect is expected to be more pronounced on a branded site than on an independent comparison site. Using an online experiment featuring four restaurant scenarios, 125 participants wrote an online review featuring sentences as variables. Results showed no significant differences for the composition of the review when the writer was either rewarded or unrewarded. Also, no significant moderating effect of platform was found. Additionally, the results did not indicate a significant direct effect of platform type. This study contributes to the eWOM literature by taking the writers’
perspective and examining the relationship between rewards and composition of the review. Further, this study relates rewarding with the platform the review is posted on. As being the first of its kind, limitations and directions for future research are given.
Introduction
Product or service experiences are shared every day, whether it is face to face, via their mobile phone, internet or email; opinions get shared. Such interpersonal communication is known as word of mouth (Berger, 2014). The rise of the internet, and social media in
particular, have made the sharing of opinions much easier. With the internet being a medium of two-way communication, opinions can be read, created and shared with everyone (Hennig-Thurau, Gwinner, Walsh, & Gremler, 2004). With a few clicks of the mouse on your
computer you can find a multitude of other consumers’ opinions all over the world in various product- or service categories.
Due to the rise of the internet, traditional word of mouth (WOM) has expanded to the internet, which is known as electronic word of mouth (eWOM): “any positive or negative
statement made by potential, actual, or former customers about a product or company, which is made available to a multitude of people and institutions via the Internet” (Hennig-Thurau et
al., 2004, p. 39). Where in the past consumers trusted solely on traditional word of mouth from friends, families or acquainted contacts (Pan & Zhang, 2011; Zhang, Craciun, & Shin, 2010) when searching for product information, eWOM has now become the preferred method (King, Racherla, & Bush, 2014). A study by Steffes and Burgee (2009) showed that reviews from strangers do have an impact on decision making, confirming that next to strong ties (family, friends)(Brown & Reingen, 1987), weaker ties also impact decision making. This way, eWOM has become an important source of product-related information (Korfiatis, García-Bariocanal, & Sánchez-Alonso, 2012; Zhang et al., 2010). Recent research indicates that 88% of customers read online reviews to determine the quality of businesses and their products and services (Anderson, 2014). This is indicative of the importance of consumer word of mouth on purchase decisions of other consumers.
One of the strengths of eWOM is that such a review is written from the perspective of the consumer, and not the seller (Chen & Xie, 2008). Because the authors of these reviews are fellow consumers, their reviews are perceived as more credible and relevant than marketer information (Bickart & Schindler, 2001). Several studies indeed found that these online reviews, whether positive or negative, have a significant impact on product sales (Amblee & Bui, 2011; Archak, Ghose, & Ipeirotis, 2011; Chen & Xie, 2008; Chevalier & Mayzlin, 2006; Liu, 2006). As such, eWOM has empowered the consumer in the decision making process (O’Hern & Kahle, 2013) and is therefore an important factor in today’s promotion and marketing of products.
The shift of promotional power from companies to consumer has led to an increase of websites such as Amazon, where consumers can post and read online product reviews.
Acknowledging the power of eWOM, some product or service sites have begun to incorporate consumer reviews on their site (e.g. Walmart). For a review site to be successful, it is
important that consumers are willing to share information and build resources through a network of social connections, known as social capital (Coleman, 1988; Hung & Yiyan Li, 2007). To stimulate reviews, reward programs are an increasingly popular method (Verlegh, Ryu, Tuk, & Feick, 2013). Further, research has shown that stimulating reviews is more effective than traditional marketing in generating revenue, because it guides consumers in their purchase decisions (Trusov, Bucklin, & Pauwels, 2009).
As it is known that rewarding consumers to write reviews does lead to an increase in reviews (Ryu & Feick, 2007) and revenue (Trusov et al., 2009), still no evidence exists examining whether rewarding (vs no reward) changes the quality of the review. What choices do creators of eWOM make regarding the composition of reviews when receiving a reward for their efforts? Will the usefulness of the review change when the consumer is rewarded? The first aim of this study is to examine the effects of reward on composition of reviews.
Examining this issue is of relevance since there are societal concerns about the effects of rewarding eWOM/reviews. Specifically whether rewarded reviews can be rusted and used as a trustworthy and useful source of information in the purchase decision process (Willemsen, 2013).
The second aim of this study is to investigate the moderating role of platform type. In recent research, Bronner and de Hoog (2011), found different motivations for consumers posting reviews of vacations on review- and marketing sites. They found that consumers who are more self-directed —i.e. those who are also motivated by reward— write different types of reviews that other-directed consumers. Specifically, on these platforms consumers post a smaller number of reviews that cover more negative aspects of their vacation. Following from this research, it can therefore be expected that consumers which are explicitly extrinsic
motivated do post different reviews than intrinsically motivated consumers, and that there is a moderating role of platform type. Where this study investigated the influence of motivations on type of review site and motivation on message characteristics, the current study
investigates whether reviews differ in composition if the motivation for posting a review is monetary, and the influence of platform type on this effect.
So far, no research has delved into the question of how platform type (e.g. independent or branded site) would affect the effects of rewards. Examining differences in the composition of rewarded and unrewarded reviews that are posted on different platforms would shed new light on the context effects of reward programs. This leads to the following research question:
How do reward programs affect choices creators of eWOM make regarding the composition of reviews when writing online product reviews, and how are these choices affected by platform type?
Theoretical framework
Intrinsic vs extrinsic motivations
Consumers can have different motivations for writing eWOM (Hennig-Thurau et al., 2004; Muntinga, Moorman, & Smit, 2011). Poch and Martin (2014) categorize motivations in two overarching motivations: intrinsic and extrinsic motivations. Intrinsic motivations are activities which are performed without receiving a reward and where consumers are doing it for their own sake (Poch & Martin, 2014). Examples of intrinsic motivations are altruism, (dis)satisfaction with a product, or loyalty (Hennig-Thurau et al., 2004; Muntinga et al., 2011; Poch & Martin, 2014). Extrinsic motivations are focused on receiving a form of reward for the effort or creating eWOM. Where intrinsic motivations are focused on the activity itself, extrinsic motivations are focused on the reward (Poch & Martin, 2014). When an individual is rewarded for their effort to create an online review, the goal of the individual is to persuade other consumers to like or dislike a product (Kim, Naylor, Sivadas, & Sugumaran, 2015). Examples of extrinsic motives are social benefits and economic rewards (Hennig-Thurau et al., 2004; Poch & Martin, 2014).
Rewarding consumers for their effort of creating online reviews instead of relying on intrinsic motives can have positive outcomes for marketers. Several studies showed that rewarding increases the quantity of reviews (Ryu & Feick, 2007), provides other consumers with balanced and informative feedback (King et al., 2014), increases revenue (Trusov et al., 2009) and prevent an under-reporting bias (Chevalier & Mayzlin, 2006; King et al., 2014). The under-reporting bias suggests that only extremely satisfied and unsatisfied customers write reviews and thus create a skewed picture (Chevalier & Mayzlin, 2006; King et al., 2014). In order to achieve this, organizations can provide different types of rewards for consumers, such as monetary rewards, discounts, prizes or job-related rewards (Hennig-Thurau et al., 2004; Muntinga et al., 2011).
Wang, Teo and Wei (2009) describe that an individual, in order to contribute an online review, has to be attracted by the extrinsic motive —i.e. a monetary reward— that is offered. Further, they judge the probability of obtaining the reward given their own product
knowledge and consumption experience, which they call assessment of goal pursuit. This assessment of goal pursuit shows that the reward has to be substantially enough to evoke differences in choices and contributing altogether. A study by Verlegh, de Vries, Willemsen and Schulte (2012) found that when the economic reward for writing an online review is not substantial enough, no difference in composition of the online review between rewarded and unrewarded reviews is found.
Although previous research have found a direct link between rewards and eWOM behaviour, less is known about the way the eWOM message is formulated. When exploring the composition of the rewarded online reviews, research has shown that communicators of eWOM tailor their message to the audience (Kim et al., 2015). The tailoring of the message is tied to the motivations of the consumer. In this study the composition of online reviews for extrinsic motivations are explored, which translate to the reward a consumer gets for writing an online review (Kim et al., 2015). This study sets out to examine the effects of reward on composition of online reviews, with focus on rating, type of arguments, review length, valence and punctuation cues. Further, the moderating role of platform type is discussed.
The effects of extrinsic rewards on the composition of the review
Rating
Online consumers reviews contain ratings and open-ended text (Park & Kim, 2008). Ratings are numeric statistics, mostly featuring a star-rating (e.g. one to five, one to seven), capturing the reviewers assessment of the product or brand (Willemsen, Neijens, Bronner, & de Ridder, 2011). Various research has shown that ratings have a significant impact on sales
and can even predict future sales (Chevalier & Mayzlin, 2006; Clemons, Gao, & Hitt, 2006), with an improvement in rating leading to an improvement in sales (Chevalier & Mayzlin, 2006). Further, they find that low ratings have a stronger impact on sales than high ratings.
An explanation for the increased sales of high-rated products would involve the abundance of high-rated products on websites, for which Chevalier and Mayzlin (2006) found that more than 50 percent of the book reviews on Amazon.com and over 65 percent of the book reviews on barnesandnoble.com are five-star reviews. Basis for this prevalence could be a confirmatory basis, where consumers who create online reviews justify or defend their purchase with high-rated reviews (Mynatt, Doherty, & Tweney, 1977; Qiu, Pang, & Lim, 2012). A second explanation for increased sales due to high-rated reviews could be that because of the high number of high-rated reviews, the product or service is considered popular, and thus displayed more prominently on the site. Evidence for the effect of volume of online reviews on sales is found by Duan, Gu, and Whinston (2008), who investigated the effect of online reviews on offline box office. They postulate that increased awareness due to the large volume of online reviews can lead to higher sales.
Turning towards the composition of online reviews, the question arises whether rewarded online reviews would differ in (average) rating from unrewarded reviews. Based on the “saying is believing”-effect (Kim et al., 2015), it can be argued that in order to persuade the audience of their opinion on the product, scoring of the online review could be more extreme when receiving a reward compared to no reward. The “saying is believing”-effect suggests that people tailor their message to their audience (Kim et al., 2015). Receivers of rewards will be more likely to try and convince the audience than consumers who have intrinsic motives. This is demonstrated by a study by Verlegh et al. (2012), who found that reviews posted in a period with a substantial reward were significantly more extreme than reviews posted in a period with small or no rewards. While the average rating did not change
for rewarded and unrewarded conditions, the ratings did become more extreme. Berger (2014) backs the argumentation by Kim et al. (2015) and argues that when the goal of the writer of eWOM is to convince someone that something is good or bad, people should share more extreme positive or negative information (rather than moderately positive or negative). Based on this research, it is expected that consumers who are rewarded for submitting an online review will provide a more extreme rating than consumers who do not get a reward.
H1a: Rewarded reviews contain more extreme ratings that no-rewarded reviews
Arguments
As mentioned before, next to (star) ratings, online reviews do consist of open ended text where consumers have the opportunity to fill out the pros and cons of the product or service they are reviewing. It is generally assumed that more argumentation has more persuasive capabilities (Petty & Cacioppo, 1986), and make other people more likely to comply with the message (Willemsen et al., 2011). Aristoteles mentioned several thousands of years back in his book Ars Rhetorica that there are three technical means of persuasion, namely logos, ethos and pathos (Braet, 1992). Logos, his concept of argumentation from which the word logic is derived, can be used as a way of persuading the audience with objective arguments (Braet, 1992). Secondly, ethos is considered as an ethical proof of speaker, where the goal is to convince the audience that the speaker is a credible person. Aristotle defined the true credibility of the speaker as a combination of the qualities good sense, virtue and goodwill attributed to the speaker by the audience (Braet, 1992). Aristotle defined the third form of convincing an audience —pathos— as “putting the hearer into a certain frame of mind” (Braet, 1992, p. 314). Pathos is the use of emotions in a persuasive way in order to make the audience like the speaker and influence their judgment (Braet, 1992).
In an ideal situation, a fusion of ethos, pathos and logos is considered to be the most persuasive to an audience (Braet, 1992). This would especially hold in an online environment, where consumers cannot rely on social cues, such as physical presence, facial expressions and gestures, which are present in a face to face conversation (Walther, 1996). In the absence of those social cues, consumers use argumentation as a cue to assess the usefulness of a review (Willemsen et al., 2011) and the perceived credibility of the writer (Ballantine & Cheung, 2015).
A study by Otterbacher (2011) found that logos in reviews is directly correlated with the value of those reviews; more logos is found in more valuable reviews (displayed on the front page) compared to less valuable reviews (displayed on the middle and last page). Similar effects were reported in a content analysis by Willemsen et al. (2011). In their study,
containing 400 online reviews, Willemsen et al. (2011) found that an important content characteristic of an online review for it to be perceived as useful was argument density. The higher the density of argumentation, the more useful an online review was experienced. Similar effects were reported by Ballantine and Cheung (2015), who found that objective comments in online reviews have a positive influence on perceived source credibility, which is consistent with previous research (Doh & Hwang, 2008).
When specifically looking at differences in argumentation volume for ethos, pathos and logos in rewarded and unrewarded conditions, research shows that vacationers who primarily have self-directed motivations, such as financial gain by reward, mention less rational arguments about their vacation in an online review than vacationers who have more other-directed motivations (Bronner & Hoog, 2011). As arguments are an important element of persuasion techniques, it is striking that consumer with more self-directed motivations mention less aspects of their vacation. A study by Verlegh et al. (2012) compared incentivized conditions and found that when there was a large reward (vs small and no reward), consumers
wrote reviews which made significantly less use of logos (rational arguments). Based on the studies by Bronner and de Hoog (2011) and Verlegh et al. (2012), it is expected that in the rewarded condition consumers will post less rational arguments than consumers in the unrewarded condition.
H1b: Rewarded reviews contain less rational arguments than no reward reviews.
Review length
The length of online reviews has been shown to impact the appreciation of the review by consumers (Chevalier & Mayzlin, 2006) and on the helpfulness of the review (Pan & Zhang, 2011). Chevalier and Mayzlin (2006) found that review length can be correlated with star ratings, as ratings greater than 4 (out of 5) for books have less characters than reviews with less than 4 stars. They argue that this could be due to including both positive and negative aspects in the review, which makes it lengthier than reviews who are only very positive or very negative.
In this study, we focus on the effect of reward on review length. Other research, by Verlegh et al. (2012), found that consumers who received a large reward for their review wrote significantly more words than consumers who received a small or no reward. This result is in contrast with the findings of Bronner and de Hoog (2011), who find in their study that vacationers with more self-directed motivations such as financial gain or venting write less text than consumers who are more other-directed. Combining this information with the number of arguments, which is lower in the rewarded condition, it is expected that consumers will write less longer reviews in the rewarded condition than in the no reward condition.
Valence
One of the most important review characteristics of the composition of an online review is valence. Valence refers to the ratio of negative and positive information in the written part of an online review as created by consumers (King et al., 2014). Prior research into the influence of eWOM valence on consumers’ behaviour and outcome variables revealed mixed results (Ballantine & Cheung, 2015), with some studies found that negative information has more impact on sales than positive information (Chevalier & Mayzlin, 2006). Berger, Sorensen, and Rasmussen (2010) found that negative publicity increased sales by raising awareness. Other studies argued that positive information has more impact due to the larger amount of positive reviews (Sorensen & Rasmussen, 2004, Chevalier & Mayzlin, 2006), whereas some find that mixed information (both positive and negative) is most
influential (Vermeulen & Seegers, 2009; Willemsen et al., 2011). Valence of product reviews is also known to moderate consumers’ brand attitudes (Lee, Rodgers, & Kim, 2009). While valence is identified as an important variable for eWOM, it is still unclear what motivates consumers to choose a certain valence in different conditions.
A study by Verlegh et al. (2012) used a content analysis to find differences in
composition of online reviews between rewarded conditions (large reward, small reward and no reward). They found no difference in valence between those conditions, either more negative or positive. On the other hand, survey research by Bronner and de Hoog (2011) found that consumers who have more self-directed motivations, such as rewards, are more negative in tone than consumers who have other-directed motivations, such as help for others. While this study explored self-directed motivations such as rewards, other motivations
(empowerment, venting, revenge) were included as well. The isolated effect of rewards thus remains unclear. However, since empowerment, venting and revenge and rewards are all persuasion motives, the isolated effect of the latter motive is expected to be similar to the
effects of the other motives. Hence, based on Bronner and de Hoog (2011), it is expected that the tone in online reviews will be more negative when consumers are rewarded compared to the no reward condition. This leads to the following hypothesis:
H1d: In the reward condition, writers of eWOM are more negative in tone than in the no reward condition.
Punctuation cues
Online reviews are textual statements shared via computer-mediated communication (Hennig-Thurau et al., 2004), rather than face to face (Pan & Zhang, 2011). In the online environment there is a lack of social cues or richness, since it is not possible to transfer facial expressions and emotions in an review (Otterbacher, 2011). This is known as media richness, which is the ability to facilitate communication within a time interval to achieve mutual understanding (Sun & Cheng, 2007). Face to face communication is the highest form of media richness because it provides immediate feedback, features multiple cues (speech, gestures and facial expressions), has a personal focus (such as emotions and feelings) and can feature language variety (natural language instead of only numbers and formulas) (Sun & Cheng, 2007). On the other hand, computer-mediated communication, such as online reviews, are low in richness because of the lack of immediate feedback, less cues and less personal focus. To compensate for this lack of richness, writers of online reviews use textual features such as nonstandard use of capitalized letters (all letters capitals) or punctuation marks (e.g. !! or ??) (Otterbacher, 2011).
Research into textual features for online reviews found that reviews which are more valuable (displayed on the front page) contained less punctuation marks than less valuable reviews (displayed on the middle and last page) (Otterbacher, 2011). Thus it seems that the absence of these textual features is perceived more effective. The only study found for
rewarded conditions confirmed findings by Otterbacher (2011) and expanded these findings to rewarded conditions (Verlegh et al., 2012). In periods when there was a large reward,
consumers would write reviews which contained significantly more punctuation marks and spaces than in periods when there was a small or no reward. This difference in punctuation marks is a confirmation of the study by Ryu and Feick (2007), who found that using a reward leads to overjustification and thus more punctuation cues. Verlegh et al. (2012) found no difference between the small reward and no reward condition. Further research into textual features has yet to be performed, thus based on these findings, it is expected that in the rewarded condition consumers will use more punctuation cues than in the no reward condition.
H1e: In the reward condition, writers of eWOM use more punctuation cues than in the no reward condition.
Moderating role of platform
Consumers can choose different platform types to post their online reviews. Several studies focussing on eWOM tend to call those platforms “communities” (King et al., 2014) as an abstract concept of the platforms eWOM is posted on. Yeap, Ignatius and Ramayah (2014) find that review sites (compared to personal blogs, social networking sites and instant
messaging sites) are the most preferred platform for gathering product information. People gather on those different platforms depending on the nature with regard to either
products/services or activities/conversations (King et al., 2014). Different platform types thus seem to attract different people. In addition, Hu and Sundar (2010) argue that those different platforms account for different gatekeeping roles. Gatekeeping is the concept of selecting sources, a concept by Shoemaker, Eichholz, Kim and Wrigley (2001). Specifically, it entails “the process by which the vast array of potential news messages are winnowed, shaped, and
prodded into those few that are actually transmitted by the news media” (Shoemaker et al.,
2001, p. 233).
Specifically for online reviews, gatekeepers control or limit access to information and play the role of communication channel, link or intermediary (Laidlaw, 2010). Platform types for eWOM can also play the role of gatekeeper, as they publish content and manage the information that appears online (Laidlaw, 2010). Yeap et al. (2014) found that the credibility of the platform is the most important factor in seeking eWOM information. With perceived source credibility, consumers use prior knowledge to assess the credibility of the source, which is an element of persuasion knowledge (Friestad & Wright, 1994). When a source is perceived to have a commercial intention, consumers will activate persuasion knowledge and a change of meaning might take place which involves effects like disengaging from the interaction, draw inferences of some sort, get distracted from the message or discount what the spokesperson says (Friestad & Wright, 1994). Thus, the nature of the site influences choices consumers make for seeking and contributing eWOM information (Hu & Sundar, 2010; Yeap et al., 2014).
Research into the influence of product judgment on different platforms has been performed by Lee and Youn (2009). They compared outcome variables for, amongst other variables, three different platforms. Lee and Youn (2009) use the attribution theory to explain the impact of (e)WOM on a certain position the communicator takes regarding a service or product. Attribution theory explains how people make causal inferences and how they use that information (Kelley, 1973; Mizerski, Golden, & Kernan, 1979). People can attribute the communicator’s message to the stimulus (the product or service) and/or non-stimulus factors (characteristics of the communicator or circumstances) (Lee & Youn, 2009). Stimulus attribution is that the reader of an online review feels that the writer positively reviewed the product or service because it actually is good. Non-stimulus attribution is that when the reader
of the online review feels that the positive review is written because the writer receives an reward for the effort (Lee & Youn, 2009). For those circumstance attributions, which in the study by Lee and Youn (2009) is attributing the review not to the actual performance of the product but to the site itself, there was a significant effect for a branded website compared to an independent website and a blog. This means that the consumers in this study activated their persuasion knowledge and attributed the review to the intention of the company, which is selling the product.
In the present study platform type is manipulated so that the nature of the site (branded vs independent) changes, and thus the activation of persuasion knowledge. It is expected that consumers on branded websites recognize the nature of the site and adapt their review to the purpose of that site. Based on the previous research, it is expected that the effects of
rewarding the reviews of consumers will be more pronounced on a branded site than on an independent comparison site. This leads to the following hypothesis:
H2: The before mentioned effects of rewarding vs non-rewarding will be more pronounced in a branded site than an independent site.
Conceptual model
Based on the theory mentioned in the previous chapters, this paper sets out to find if differences in composition of an online review can be attributed to intrinsic or extrinsic conditions and what influence platform has on this relationship. The hypotheses can be summed in a conceptual model (Figure 1).
Figure 1. Conceptual model.
- Rational arguments - Review length - Valence - Punctuation cues - Rating Platform type
Intrinsic versus extrinsic motivations (Reward)
Method
Participants and Design
The hypotheses will be investigated by using an 2 (reward vs no reward condition) x 2 (independent vs brand-related platform) between groups design, with the number and type of arguments, valence, punctuation cues and extremity as dependent variables. The experiment thus consists of four experimental conditions. In total, 125 participants completed the study. The participants were between 18 and 64 years old (M = 30,66, SD = 12,59), of which 64,8% was female. The sample consisted of mostly highly educated participants (79,2% completed secondary education). See Appendix A for full overview of demographics.
Stimuli
To incorporate and manipulate both reward and platform, four scenarios were developed. The text of the scenarios was the same, except for reward and platform.
Participants were asked to imagine that they had been to a restaurant. Several positive (e.g. the food was of high quality) and less positive (e.g. the chairs were not very comfortable) experiences were mentioned in the scenario about the service, speed of service, food, atmosphere as well as the general conclusion of these experiences. The categories are based on the criteria that review sites use for evaluating restaurants and thus reflect a realistic review context. To manipulate reward type, participants were subsequently asked to imagine that they would or would not receive a reward for their review, i.e. a free bottle of wine. To manipulate platform type, participants were asked to imagine that they would post the review on either the site of the restaurant or an independent review site. All scenarios are included in Appendix D.
Procedure
The study is conducted using an online experiment, which reflects the natural context of online reviews. Participants were recruited using Facebook and email by using an enclosed web link. Participants had an incentive to win one of the two €15 vouchers in a random lottery.
When starting the questionnaire, participants were first shown the informed consent. After accepting the terms and conditions, participants were randomly assigned to one of the four conditions. Each condition featured one scenario which differed in reward (yes/no) and type of platform (independent site or the site of the restaurant). As part of the scenario, participants were asked to imagine that they had visited a restaurant. A restaurant was chosen because consumers tend to search more for eWOM on service, such as restaurants, than utilitarian products (Bei, Chen, & Widdows, 2004).
After reading the scenario, which was designed to motivate the participants intrinsically or extrinsically, participants had to choose sentences to construct their online review based on what they read in the scenario. They had eighteen sentences to choose from, with a lower limit of six and an upper limit of twelve sentences. The sentences were selected to correspond with the various independent variables and pre-tested beforehand (see Measures and Pre-test below). After selecting the sentences, participants chose a rating for the review based on a number ranging from 1 to 10. Finally, participants were asked to answer questions about the manipulation check and social demographics. They were then thanked for
participating, given the option to enter their email address to participate in the lottery and debriefed.
Measures
Participants had to choose sentences to construct their ‘online review’. Each sentence was designed to correspond with a dependent variable.
Aristotle’s ethos, logos, pathos. To strengthen the analyses and validity, measures for
ethos and pathos were included in this study, next to the measures for logos. To measure the ratio of ethos, logos and pathos in an online review, several sentences were created. For each of four categories (service, speed of service, food and décor), three sentences were employed which focused on one of the three communication tactics by Aristotle, resulting in 12
sentences in total. As these three strategies are independent, they were measured with one sentence for each. The ethos category focused on emphasizing the experience of the writer (Otterbacher, 2011) and featured sentences as “As a wine connoisseur, I appreciate this wine.”, “I regularly go to restaurants and was surprised by the different tastes that come together in the dishes.”, “In comparison with other restaurants I’ve been to, we had to wait a long time for the main dish.”. Logos focused on rational arguments, and was operationalised as “After the entrée, we had to wait forty minutes before the main dish was served.”, “The seats were made of hard wood and not very comfortable.”, “They made use of high quality seasonal products.”. The pathos sentences focused on emotional experience; “We felt very welcome.”, “Every dish was a taste explosion.”, “We became very hungry because we had to wait so long.”.
Punctuation. To measure punctuation, several sentences with repeated punctuation
marks were included in the conclusion; “If you like good food, this restaurant is
recommended!!!”, “They should replace the uncomfortable seats. If you go out for dinner, you expect to be seated properly…???”, “The good food and the ambience make me come back!!!”. To balance the options for punctuation, two similar sentences with significant less punctuation were included in the conclusion as well.
Valence & Extremity. To measure the average valence and extremity of the ratings,
participants could rate their experience with a number ranging from 1 to 10 (M = 7.40, SD = 0.62). Numeric ratings were chosen over star ratings for more elaborate measurements. Choosing the rating was performed after selecting the sentences. To calculate the extremity of the review, the ratings of the rewarded and unrewarded reviews were subtracted with the median for both conditions. The median for both conditions was 7, which means that a rating of 8 has an extremity score of 1. The average extremity score was 0.40 (SD = 0.62).
Review length. Review length was measured by the amount of sentences the
participants had chosen. They were given the option to choose between 6 and 12 (from a total of 18) sentences at their own judgment. More sentences equals longer reviews.
Manipulation check. To check if the manipulation was successful and participants
identified the reward and platform conditions correctly, two separate questions measured the manipulation. The first question asked whether the participants could remember the reward given, if any. Answers were “Yes, a free bottle of wine”, “No, there was no reward” and “I don’t remember”. For platform, the question was whether participants could remember the platform where the online review was supposed to be posted. Answers to this question were “On an independent comparison site”, “On the restaurants’ site” and “I don’t remember”.
Pre-test
A pre-test was used to assess the effectiveness of the scenarios and sentences. Using a within-subjects design (N = 11), all scenarios and sentences were tested. Participants had to read the four scenarios and were asked to recall the type of reward (free bottle of wine or no reward) and platform type the online review was supposed to be posted on (independent comparison site or restaurant site). The reward correctly was identified by 95,5% of the participants. The platform was correctly identified correctly by 84,1% of the participants. To
improve identification of the reward and platform in the scenario even more, the material was slightly adjusted by displaying more prominently in the text (text included in Appendix D).
The sentences were pre-tested on ethos/pathos/logos, valence and punctuation. Each participant had to rate the degree to which it reflected ethos/logos/pathos, positive or negative valence and punctuation. Several sentences scored insufficient on the appropriate scale (ethos/logos/pathos) and were adjusted. Scoring of the sentences is included in Appendix E, along with the changes made. Last, in the conclusion several sentences were included which had repeated punctuation.
Results
Randomization and covariate check
A randomization check was performed to test whether the randomization created equal groups in terms of age, gender and education. An analysis of variance revealed no significant differences for age between the four experimental groups F(3, 121) = 0.62, p = 0.61.
Additionally, a crosstab analysis showed no significant differences for gender (χ2(3) = 4.75,
p = .19) and education (χ2(15) = 8.57, p = .90). As a result, the randomization can be considered successful.
A Bivariate Pearson correlation test was performed to check for possible alternate explanations. The correlation analysis tested for relationships between the dependent variables and the control variables age, gender and education. Results showed that age correlated
positively with the dependent variable rating (r = -0.19, N = 125, p = 0.03). Age was thus inserted in the appropriate analysis as a covariate. The complete correlation results are in the Appendix B.
Manipulation check
Results for the manipulation check showed that 84% of the participants (n = 105) did identify the reward from the scenario. The manipulation check for the platform revealed that 81,6% of the participants (n = 102) did identify the platform from the scenario. Participants who did not identify the reward or platform correct were removed from appropriate analyses.
Hypotheses Testing
To test the hypotheses, a two-way Multivariate Analysis of Covariance (MANCOVA) was conducted with reward and type of platform as independent variables, age as covariate, and rating, rating extremity, logos, ethos, pathos, review length, and punctuation cues as dependent variables. The results are discussed below.
The effects of rewards. The MANCOVA showed no significant main effect for
rewards (Wilks’ Lambda=0.95, F (6,83)= 0.710, p=.64, ŋ2=.05). This means that rewarded reviews did not yield difference from unrewarded reviews in terms of rating, rating extremity, logos, ethos, pathos, review length and punctuation cues. Thus, hypotheses 1a-1e were not supported.
The effects of platform type. No specific effects were expected for platform, and the
MANCOVA showed no significant main effect for platform type (Wilks’ Lambda=0.94,
F (6,83)= 0.824, p=.56, ŋ2=.06). This means that the independent comparison site did not
yield difference from the restaurant site in terms of rating, rating extremity, logos, ethos, pathos, review length and punctuation cues.
Interaction effect. The MANCOVA showed no significant interaction effect for
reward and platform type (Wilks’ Lambda=0.99, F (6,83)= 0.131, p=.99, ŋ2=.01). This means that there is no significant combined effect of reward and platform type on the dependent
variables rating, rating extremity, logos, ethos, pathos, review length and punctuation cues. Thus, hypothesis 2 is not supported.
Conclusion and discussion
This paper set out to examine the effect of rewards on the composition of online reviews, and whether platform type influenced that effect. This was investigated for rewarded and unrewarded conditions, with the moderating effect of platform using the main research question: “How do reward programs affect choices creators of eWOM make regarding the
composition of reviews when writing online product reviews, and how are these choices affected by platform type?”.
Based on previous research, it was expected that rewarding consumers would make them compose an online review with more punctuation, less rational arguments, more extreme ratings, more negative valence and shorter reviews compared to unrewarded reviews. It was hypothesized that this effect was more pronounced when the review was posted on a branded site compared to an independent website. To test the hypotheses, an online experiment was conducted in which reward and platform were manipulated in four scenarios.
This study set out with two aims, with the first being to examine the effects of reward on composition of reviews. The results of this experiment did not support the hypothesized effect of reward on any of the dependent variables. There were no significant differences between rewarded and unrewarded reviews for extremity of ratings, Aristotle’s ethos, logos and pathos, review length, valence and punctuation cues. It can be assumed, based on these results, that there is no difference in the composition of a review between consumers who are extrinsically and intrinsically motivated. Thus, it can be argued that consumers in both conditions use the same amount of personal information, emotional appeal and rational arguments for convincing their readers. Further, reviews that are intrinsic and extrinsic
motivated do not differ in the extremity of the rating of the review, length, valence and punctuation cues.
The second aim in this experiment was to investigate hypothesized moderating effect of platform type on the relationship between reward and the dependent variables. It was argued that when the review would be posted on a branded site, the before mentioned effects of rewarding vs non-rewarding would be more pronounced. The results do not support the hypothesized effects. There is no significant difference between posting the review on an independent website or on a branded site for Aristotle’s ethos and logos, extremity of the review, valence and punctuation cues.
Concluding the present study, and answering the research question, rewarding the consumer does not change the composition of an online review compared to unrewarded consumers. This effect is not significantly moderated by the platform type the online review is posted on. While the hypothesized findings are not significant, these results contribute to the knowledge of eWOM and can be used to guide further research.
Implication of the findings
The present study is the first to examine the effects of rewarding for different types of platform, but the results did not yield a significant effect for rewarding on the composition of an online review. Further, no significant moderating effect of platform is found. However, the findings of this study have positive implications for both consumers and marketers. First, the lack of difference in the composition between rewarded and unrewarded reviews shows that that consumers are not being misled by rewarded reviews. With the increasing use of rewards by marketers to improve the quantity of online reviews (Ryu & Feick, 2007; Verlegh et al., 2013), trustworthiness of online reviews is a valid concern because consumers favour user-generated reviews over expert reviews for high volumes (Flanagin & Metzger, 2013). This is
in contrast with previous research, which has suggested that the consumers motivations to write an online review influences the composition of the review (Yap, Soetarto, & Sweeney, 2013, Verlegh et al., 2012). However, the results in this study indicate that the differences in composition of the online review between intrinsic and extrinsic conditions are too small to be of significance. Furthermore, the findings are positive for marketers because rewarding
consumers for writing online reviews does not result in a different composition of the reviews, which means that marketers can use rewards to increase the quantity of online reviews.
Second, this study is the first to examine the effect of rewarding for different types of platform, for which no moderating or direct effect of platform type on the composition of online reviews was found. This means that there is no difference between online reviews on branded or independent comparison sites. The results are in contrast with the findings by Bronner and de Hoog, who found that there is difference in composition of online reviews on marketer and independent comparison sites (Bronner & de Hoog, 2010; Bronner & de Hoog, 2011). The results of the current study are positive for both marketers and consumers, as consumers are not being misled on branded sites and marketers can use their own platform type for online reviews.
Limitations and future research
Although the present paper adds to the knowledge of eWOM motivation and
composition of online consumers reviews, some limitations are to be taken into account. One possible explanation for the differential findings could be that participants are already
extrinsically or intrinsically motivated before the start of the experiment. Both the willingness to help the researcher and the chance to win one of the two vouchers as a reward could be responsible for having an effect on the results. The latter alternative explanation can be confirmed by the large amount of email addresses entered in the lottery (95). As participants
are then already intrinsically or extrinsically motivated, depending on whether they participated to aid the researcher or to win the vouchers, the manipulation might not be as effective as intended. An idea would be to combine a real life setting with survey research. Using an existing platform for online reviews and combining this with a survey after posting the review, motivations can be gathered. When combining the data from the survey with a content analysis on the reviews posted, a complete picture can be painted. An example could be the study by Verlegh et al. (2012), who cooperated with a real eWOM platform provider and used a content analysis to compare data between rewarded conditions. Combining different approaches in a real life setting, the influence of pre-existing motivations will be reduced. Further, with the use of a real life setting with a content analysis, the possible limitation of the pre-made sentences will be solved and validity improved.
A second possible limitation is the size of the reward. In the present study a free bottle of wine was offered in the rewarded conditions, and no reward in the intrinsic motivated conditions. A similar study by Verlegh et al. (2012) found no significant difference between in the content of the review between no reward and a small reward, but did find differences in content for the large reward. Wang et al. (2009) underlined that the reward has to be
substantial enough to evoke different consumer responses. Possibly the reward in the present study was not significant enough to evoke differences in the choices of the participants. For future research, it is recommended that when rewarded conditions are compared, the size of the reward is substantial enough.
Last, a possible explanation for the lack of difference between the conditions in this study could be the personality traits of the participants, such as the interdependent
self-construal. The interdependent self-construal refers to how consumers perceive themselves and the relation to others around them in their community (Ahluwalia, 2008; Lee, Kim, & Kim, 2012). Two types of self-construal have been identified, the independent —those who value
their uniqueness and individuality— and interdependent —those who value being connected to a group— self construal. Several studies have found differences in eWOM behaviour (Lee et al., 2012, Wang, Yang, & Wang, 2014). In the present study, no check for self-construal was imbedded, thus future research should take the self-construal in consideration.
While the results are not significant and several limitations mentioned in this chapter, the present study does add to the body of eWOM-literature. This study uses a method which has never been attempted to try and create a natural environment for writing an online review. By examining intrinsic and extrinsic motivations combined with the moderating effect of platform type, this study brings together the message and the medium in eWOM territory.
APPENDIX A
Table 1. Demographics (N = 125).
Gender n % Age Educational
level
n %
Female 81 64,8 Minimum 18 Primary school 1 0,8
Male 44 35,2 Maximum 64 High school 4 3,2
MBO 21 16,8
Secondary education
APPENDIX B Table 2: Bivariate product moment correlation matrix.
Variables 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Ethos 1 2. Logos -.531** 1 3. Pathos -.164 -.433** 1 4. Punctuation .038 .085 .029 1 5. Nr. Arguments .379** .160 .309** .186* 1 6. Rating .026 .078 -.031 .201* 098 1 7. Service (D) .183* .047 .042 -.012 .330** -.053 1 8. Speed of Service (D) .240* .043 .260** -.039 .627** -.015 .100 1 9. Décor (D) .050 .080 .297** .181* .479** .324** .016 .060 1 10. Food (D) .315** .150 .063 .212** .646** -.045 -.044 .154 -.018 1 11. Age -.125 .014 -.012 .097 -.149 .191* -.102 -.119 .046 -.139 1 12. Gender .018 .008 -.103 -.129 -.079 -.145 -.107 .001 -.061 -.032 .084 1 13. Education -.168 .147 -.081 -.082 -.105 -.067 -.138 -.097 -.081 .033 -.477** -.144 1 Note: N = 125. D = dummy variable.
* Correlation is significant at .05 level ** Correlation is significant at .01 level
APPENDIX C
Questionnaire
Figure 7. Sentences.
Figure 9. Demographics.
APPENDIX D
Scenarios
Reward / independent comparison site
Scenario
In this study I would like you to imagine that you are a fan of good food and that you like to go out for dinner.
Recently you visited a new restaurant. After visiting the restaurant the restaurant owner asks if you want to write a review about your visit.
You remember the following points:
Service: You were welcomed with an aperitif and an appetizer, after which your order was taken (within 10 minutes).
Food: The food was of very high quality. All meals were tasty, prepared with high quality seasonal products.
You did have to wait long between the entrée and the main course (40 minutes). Décor: The restaurant had a modern interior (with high tables made from hard wood,
and the same material stools), which weren’t very comfortable. This was a bit annoying.
Task
In exchange for writing a review, you will get a free bottle of wine.
The message states that the review is to be placed on an independent comparison site.
You decide to write the review and to place it on an independent comparison site to earn the free bottle of wine.
On the next page you will see 18 sentences which can be used to write a review.
Pick the sentences you want to include in your review. You can choose a minimum of 6 and a maximum of 12 sentences. There are no wrong answers, I am only interested in how you would write your review.
Reward / restaurant site
Scenario
In this study I would like you to imagine that you are a fan of good food and that you like to go out for dinner.
Recently you visited a new restaurant. After visiting the restaurant the restaurant owner asks if you want to write a review about your visit.
You remember the following points:
Service: You were welcomed with an aperitif and an appetizer, after which your order was taken (within 10 minutes).
Food: The food was of very high quality. All meals were tasty, prepared with high quality seasonal products.
You did have to wait long between the entrée and the main course (40 minutes). Décor: The restaurant had a modern interior (with high tables made from hard wood,
and the same material stools), which weren’t very comfortable. This was a bit annoying.
Task
In exchange for writing a review, you will get a free bottle of wine.
The message states that the review is to be placed on the site of the restaurant.
You decide to write the review and to place it on the site of the restaurant to earn the free bottle of wine.
On the next page you will see 18 sentences which can be used to write a review.
Pick the sentences you want to include in your review. You can choose a minimum of 6 and a maximum of 12 sentences. There are no wrong answers, I am only interested in how you would write your review.
No reward / independent comparison site
Scenario
In this study I would like you to imagine that you are a fan of good food and that you like to go out for dinner.
Recently you visited a new restaurant. After visiting the restaurant you decide to write a review about your visit.
You remember the following points:
Service: You were welcomed with an aperitif and an appetizer, after which your order was taken (within 10 minutes).
Food: The food was of very high quality. All meals were tasty, prepared with high quality seasonal products.
You did have to wait long between the entrée and the main course (40 minutes). Décor: The restaurant had a modern interior (with high tables made from hard wood,
and the same material stools), which weren’t very comfortable. This was a bit annoying.
Task
You decide to place the review on an independent comparison site.
On the next page you will see 18 sentences which can be used to write a review.
Pick the sentences you want to include in your review. You can choose a minimum of 6 and a maximum of 12 sentences. There are no wrong answers, I am only interested in how you would write your review.
No reward / restaurant site
Scenario
In this study I would like you to imagine that you are a fan of good food and that you like to go out for dinner.
Recently you visited a new restaurant. After visiting the restaurant you decide to write a review about your visit.
You remember the following points:
Service: You were welcomed with an aperitif and an appetizer, after which your order was taken (within 10 minutes).
Food: The food was of very high quality. All meals were tasty, prepared with high quality seasonal products.
You did have to wait long between the entrée and the main course (40 minutes). Décor: The restaurant had a modern interior (with high tables made from hard wood,
and the same material stools), which weren’t very comfortable. This was a bit annoying.
Task
You decide to place the review on the site of the restaurant.
On the next page you will see 18 sentences which can be used to write a review.
Pick the sentences you want to include in your review. You can choose a minimum of 6 and a maximum of 12 sentences. There are no wrong answers, I am only interested in how you would write your review.
APPENDIX E Pre-test results Table 1. Pre-test results.
Sentence Ethos Logos Pathos Valence Punctuation
Als wijnliefhebber kan ik dit erg waarderen. (ethos)
M = 4,18 SD = 2,23 ab M = 2,64 SD = 1,86 a M = 5,36 SD = 1,50 b M = 5,82 SD = 0,60 M = 1,82 SD = 1,40 We kregen een wijn naar keus
met kleine aperitiefhapjes erbij. (logos) M = 3,36 SD = 2,01 b M = 6,27 SD = 0,91 a M = 2,82 SD = 1,47 b M = 4,73 SD = 0,91 M = 1,91 SD = 1,38 We voelden ons meteen erg
welkom. (pathos) M = 3,09 SD = 1,58 a M = 2,09 SD = 1,04 a M = 6,00 SD = 1,18 b M = 6,27 SD = 0,47 M = 1,91 SD = 1,38 In vergelijking tot andere
restaurants, moesten we erg lang wachten op het hoofdgerecht. (ethos) M = 5,55 SD = 0,82 b M = 4,18 SD = 1,89 a M = 4,36 SD = 2,11 ab(0,07) M = 2,45 SD = 0,82 M = 1,91 SD = 1,38
Na het voorgerecht duurde het 40 minuten voordat het hoofdgerecht werd geserveerd. (logos) M = 3,18 SD = 1,83 b M = 6,64 SD = 0,51 a M = 2,73 SD = 1,79 b M = 3,18 SD = 0,87 M = 1,91 SD = 1,38
We werden erg hongerig van het lange wachten. (pathos)
M = 3,55 SD = 1,51 a M = 3,36 SD = 1,96 a M = 5,55 SD = 1,37 b M = 2,27 SD = 1,27 M = 2,00 SD = 1,41 Ik ga met enige regelmaat uit
eten, maar was positief verrast door de wijze waarop smaken samenkwamen. (ethos) M = 6,00 SD = 0,89 b M = 3,18 SD = 1,78 a M = 6,09 SD = 1,04 b M = 5,73 SD = 1,68 M = 1,91 SD = 1,38
Je kunt proeven dat er alleen gebruik wordt gemaakt van kwalitatief goede seizoensproducten. (logos) M = 4,82 SD = 1,40 a M = 4,00 SD = 1,61 a M = 6,00 SD = 1,10 b M = 5,36 SD = 0,92 M = 1,91 SD = 1,38
Elk gerecht was een smaakexplosie. (pathos) M = 3,27 SD = 1,35 a M = 2,27 SD = 1,27 a M = 6,45 SD = 0,69 b M = 5,36 SD = 1,12 M = 1,91 SD = 1,38 Voor iemand die graag dineert
is dit geen pretje. (ethos)
M = 4,27 SD = 1,68 b M = 2,36 SD = 1,12 a M = 5,27 SD = 1,79 b M = 2,09 SD = 0,70 M = 1,91 SD = 1,38 De hoge tafels van hard hout
en dito krukken waren hier verantwoordelijk voor. (logos)
M = 3,82 SD = 1,54 a M = 5,18 SD = 1,78 a M = 5,00 SD = 1,61 a M = 2,82 SD = 0,98 M = 1,91 SD = 1,38 We verlieten het restaurant met
houten billen. (pathos)
M = 3,45 SD = 1,37 b M = 2,09 SD = 0,94 a M = 6,18 SD = 0,98 c M = 1,82 SD = 0,75 M = 1,91 SD = 1,38 Als je van lekker eten houd, is
dit restaurant een aanrader!!!
M = 3,82 SD = 1,60 a M = 2,91 SD = 1,58 a M = 5,27 SD = 1,62 b M = 6,09 SD = 0,70 M = 4,91 SD = 2,21 Ik kom hier graag nog eens
terug. M = 3,27 SD = 1,56 a M = 3,64 SD = 1,75 a M = 5,27 SD = 1,68 b M = 5,82 SD = 0,87 M = 1,91 SD = 1,38 Ze mogen wel wat meer
aandacht besteden aan comfort.
M = 4,00 SD = 1,55 a M = 2,82 SD = 1,25 a M = 5,36 SD = 1,03 b M = 2,27 SD = 0,79 M = 1,91 SD = 1,38 De ongemakkelijke stoelen
mogen vervangen worden. Als je uiteten gaat wil je toch lekker kunnen zitten…????
M = 3,45 SD = 1,21 a M = 3,09 SD = 1,45 a M = 5,82 SD = 1,17 b M = 2,55 SD = 1,44 M = 5,27 SD = 1,90
Table 2. Manipulation check pre-test.
Scenario Reward correct Platform correct
Reward / Independent site 100% 100%
Reward / Restaurant site 100% 72,7%
No reward / Independent site 81,8% 72,7% No reward / Restaurant site 100% 90,9%
Table 3. Changed sentences.
Sentence (old) Sentence (new)
Als wijnliefhebber kan ik dit erg waarderen. (ethos)
Als wijnkenner kan ik dit waarderen.
Ik ga met enige regelmaat uit eten, maar was positief verrast door de wijze waarop smaken samenkwamen. (ethos)
Ik ga regelmatig uit eten en was verrast door de verschillende smaken.
Je kunt proeven dat er alleen gebruik wordt gemaakt van kwalitatief goede seizoensproducten. (logos)
Er werd gebruik gemaakt van kwalitatief goede seizoensproducten.
Voor iemand die graag dineert is dit geen pretje. (ethos)
Deze stoelen zitten minder goed dan in andere restaurants waar ik geweest ben.
De hoge tafels van hard hout en dito krukken waren hier verantwoordelijk voor. (logos)
De hoge tafels en krukken waren van hard hout en niet zo prettig.
Extra punctuation: Het goede eten en de sfeer zorgen dat ik wel terug kom!!!
Extra punctuation: Als ze die stoelen vervangen krijgen ze een hoger cijfer!!!
Literature
Ahluwalia, R. (2008). How Far Can a Brand Stretch? Understanding the Role of Self-Construal. Journal of Marketing Research, 45(3), 337–350.
Amblee, N., & Bui, T. (2011). Harnessing the Influence of Social Proof in Online Shopping: The Effect of Electronic Word of Mouth on Sales of Digital Microproducts.
International Journal of Electronic Commerce, 16(2), 91–114.
doi:10.2753/JEC1086-4415160205
Anderson, M. (2014). 88% Of Consumers Trust Online Reviews As Much As Personal Recommendations. Search Engine Land. Retrieved on March 10, 2015 from http://searchengineland.com/88-consumers-trust-online-reviews-much-personal-recommendations-195803
Archak, N., Ghose, A., & Ipeirotis, P. G. (2011). Deriving the Pricing Power of Product Features by Mining Consumer Reviews. Management Science, 57(8), 1485–1509. doi:10.1287/mnsc.1110.1370
Ballantine, P. W., & Au Yeung, C. (2015). The effects of review valence in organic versus sponsored blog sites on perceived credibility, brand attitude, and behavioural
intentions. Marketing Intelligence & Planning, 33(4), 508–521. doi:10.1108/MIP-03-2014-0044
Bei, L.-T., Chen, E. Y. I., & Widdows, R. (2004). Consumers’ Online Information Search Behavior and the Phenomenon of Search vs. Experience Products. Journal of Family
and Economic Issues, 25(4), 449–467. doi:10.1007/s10834-004-5490-0
Berger, J. (2014). Word of mouth and interpersonal communication: A review and directions for future research. Journal of Consumer Psychology, 24(4), 586–607.
Berger, J., Sorensen, A. T., & Rasmussen, S. J. (2010). Positive Effects of Negative Publicity: When Negative Reviews Increase Sales. Marketing Science, 29(5), 815–827.
doi:10.1287/mksc.1090.0557
Bickart, B., & Schindler, R. M. (2001). Internet forums as influential sources of consumer information. Journal of Interactive Marketing, 15(3), 31–40. doi:10.1002/dir.1014 Braet, A. C. (1992). Ethos, pathos and logos in Aristotle’s Rhetoric: A re-examination.
Argumentation, 6(3), 307–320. doi:10.1007/BF00154696
Bronner, F., & de Hoog, R. (2010). Consumer-generated versus marketer-generated websites in consumer decision making. International Journal of Market Research, 52(2), 231– 248.
Bronner, F., & Hoog, R. de. (2011). Vacationers and eWOM: Who Posts, and Why, Where, and What? Journal of Travel Research, 50(1), 15–26. doi:10.1177/0047287509355324 Brown, J. J., & Reingen, P. H. (1987). Social Ties and Word-of-Mouth Referral Behavior.
Journal of Consumer Research, 14(3), 350–362.
Chen, Y., & Xie, J. (2008). Online Consumer Review: Word-of-Mouth as a New Element of Marketing Communication Mix. Management Science, 54(3), 477–491.
doi:10.1287/mnsc.1070.0810
Chevalier, J. A., & Mayzlin, D. (2006). The Effect of Word of Mouth on Sales: Online Book Reviews. Journal of Marketing Research, 43(3), 345–354. doi:10.1509/jmkr.43.3.345 Clemons, E. K., Gao, G. G., & Hitt, L. M. (2006). When Online Reviews Meet
Hyperdifferentiation: A Study of the Craft Beer Industry. Journal of Management
Information Systems, 23(2), 149–171. doi:10.2753/MIS0742-1222230207
Coleman, J. S. (1988). Social Capital in the Creation of Human Capital. American Journal of
Doh, S.-J., & Hwang, J.-S. (2008). How Consumers Evaluate eWOM (Electronic Word-of-Mouth) Messages. CyberPsychology & Behavior, 12(2), 193–197.
doi:10.1089/cpb.2008.0109
Duan, W., Gu, B., & Whinston, A. B. (2008). The dynamics of online word-of-mouth and product sales—An empirical investigation of the movie industry. Journal of Retailing,
84(2), 233–242. doi:10.1016/j.jretai.2008.04.005
Flanagin, A. J., & Metzger, M. J. (2013). Trusting expert- versus user-generated ratings online: The role of information volume, valence, and consumer characteristics.
Computers in Human Behavior, 29(4), 1626–1634. doi:10.1016/j.chb.2013.02.001
Friestad, M., & Wright, P. (1994). The Persuasion Knowledge Model: How People Cope with Persuasion Attempts. Journal of Consumer Research, 21(1), 1. doi:10.1086/209380 Hennig-Thurau, T., Gwinner, K. P., Walsh, G., & Gremler, D. D. (2004). Electronic
word-of-mouth via consumer-opinion platforms: What motivates consumers to articulate themselves on the Internet? Journal of Interactive Marketing, 18(1), 38–52. doi:10.1002/dir.10073
Hung, K. H., & Yiyan Li, S. (2007). The Influence of eWOM on Virtual Consumer
Communities: Social Capital, Consumer Learning, and Behavioral Outcomes. Journal
of Advertising Research, 47(4), 485–495.
Hu, Y., & Sundar, S. S. (2010). Effects of Online Health Sources on Credibility and Behavioral Intentions. Communication Research, 37(1), 105–132.
doi:10.1177/0093650209351512
Kelley, H. H. (1973). The processes of causal attribution. American Psychologist, 28(2), 107– 128.
Kim, J., Naylor, G., Sivadas, E., & Sugumaran, V. (2015). The unrealized value of
incentivized eWOM recommendations. Marketing Letters, 1–11. doi:10.1007/s11002-015-9360-3
King, R. A., Racherla, P., & Bush, V. D. (2014). What We Know and Don’t Know About Online Word-of-Mouth: A Review and Synthesis of the Literature. Journal of
Interactive Marketing, 28(3), 167–183. doi:10.1016/j.intmar.2014.02.001
Korfiatis, N., García-Bariocanal, E., & Sánchez-Alonso, S. (2012). Evaluating content quality and helpfulness of online product reviews: The interplay of review helpfulness vs. review content. Electronic Commerce Research and Applications, 11(3), 205–217. doi:10.1016/j.elerap.2011.10.003
Laidlaw, E. B. (2010). A framework for identifying Internet information gatekeepers.
International Review of Law, Computers & Technology, 24(3), 263–276.
doi:10.1080/13600869.2010.522334
Lee, D., Kim, H. S., & Kim, J. K. (2012). The role of self-construal in consumers’ electronic word of mouth (eWOM) in social networking sites: A social cognitive approach.
Computers in Human Behavior, 28(3), 1054–1062. doi:10.1016/j.chb.2012.01.009
Lee, M., Rodgers, S., & Kim, M. (2009). Effects of Valence and Extremity of eWOM on Attitude toward the Brand and Website. Journal of Current Issues & Research in
Advertising, 31(2), 1–11. doi:10.1080/10641734.2009.10505262
Lee, M., & Youn, S. (2009). Electronic word of mouth (eWOM). International Journal of
Advertising, 28(3), 473–499.
Liu, Y. (2006). Word of Mouth for Movies: Its Dynamics and Impact on Box Office Revenue.
Journal of Marketing, 70(3), 74–89.
Mizerski, R. W., Golden, L. L., & Kernan, J. B. (1979). The Attribution Process in Consumer Decision Making. Journal of Consumer Research, 6(2), 123–140.
Muntinga, D. G., Moorman, M., & Smit, E. G. (2011). Introducing COBRAs. International
Mynatt, C. R., Doherty, M. E., & Tweney, R. D. (1977). Confirmation bias in a simulated research environment: An experimental study of scientific inference. Quarterly
Journal of Experimental Psychology, 29(1), 85–95. doi:10.1080/00335557743000053
O’Hern, M. S., & Kahle, L. R. (2013). The Empowered Customer: User-Generated Content and the Future of Marketing. Global Economics and Management Review, 18(1), 22– 30. doi:10.1016/S2340-1540(13)70004-5
Otterbacher, J. (2011). Being Heard in Review Communities: Communication Tactics and Review Prominence. Journal of Computer-Mediated Communication, 16(3), 424–444. doi:10.1111/j.1083-6101.2011.01549.x
Pan, Y., & Zhang, J. Q. (2011). Born Unequal: A Study of the Helpfulness of User-Generated Product Reviews. Journal of Retailing, 87(4), 598–612.
doi:10.1016/j.jretai.2011.05.002
Park, D.-H., & Kim, S. (2008). The effects of consumer knowledge on message processing of electronic word-of-mouth via online consumer reviews. Electronic Commerce
Research and Applications, 7(4), 399–410. doi:10.1016/j.elerap.2007.12.001
Petty, R. E., & Cacioppo, J. T. (1986). The Elaboration Likelihood Model of Persuasion. In Leonard Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. Volume 19, pp. 123–205). Academic Press. Retrieved from
http://www.sciencedirect.com/science/article/pii/S0065260108602142
Poch, R., & Martin, B. (2014). Effects of intrinsic and extrinsic motivation on user-generated content. Journal of Strategic Marketing, 0(0), 1–13.
doi:10.1080/0965254X.2014.926966
Qiu, L., Pang, J., & Lim, K. H. (2012). Effects of conflicting aggregated rating on eWOM review credibility and diagnosticity: The moderating role of review valence. Decision
