Master’s Thesis
30 June 2015
Exploring Message Sidedness and
Tie-strength in
Electronic Word of Mouth
Graduate School of Communication
Master’s Programme Persuasive Communication
Author
Tomasz Jan Kawecki
Student Number
10231870
ABSTRACT
Electronic Word of Mouth (EWoM) is a relatively new development engendered by the advent of the Internet and the subsequent explosion of online activity in our modern society. In the past, traditional Word of Mouth (WoM) could result in rapid changes in consumer
perceptions and needs, as well as large spikes in sales. With a reach that spans continents and an ability to cross socio-economic barriers, EWoM represents a significant
multiplication of the effects of WoM.
A greater understanding of the determinants of consumer intention to engage in EWoM implies a deeper and richer understanding of the wider, and increasingly important, science of EWoM behavior. The purpose of this study was to add to the existing body of knowledge on the topic of intention to engage in EWoM by determining the extent to which message sidedness in online reviews has an impact on intention to engage in EWoM and what the role of tie-strength is in moderating intention to engage in EWoM.
Although the research findings were inconclusive, the research has implications for future research. For instance, a follow-up study that eliminates the limitations that the research described in this paper faced would be a worthwhile contribution to the body of knowledge on the topic of EWoM, since it would open the door to further research into other determinants, such as product compatibility with consumer need, prior knowledge of the product and visual cues relating the product.
1. Introduction and Research Question
Why do people interact? Because they do not want to feel alone? Because they cannot define their own views without the benchmarks that interaction gives them? Are they unsure of their own assessments without the feedback of others? Or perhaps they want to know who they are, who they are going to be and what their place in society is? Interaction, and the receipt or delivery of information during interaction, shapes the people involved and gives them
knowledge that helps them to justify their opinions and perceptions. Perceptions are at the basis of our personality, our views, our beliefs, our self-esteem and our personality.
People love to share information that they think is important, useful and/or emotional (Berger, 2013). The author explains that our surroundings are a reflection of things, behaviors and norms that ‘catch on’ and spread broadly to create a visible picture of reality. A first name that is very popular for newborns in one country or geographic region, for instance, may be very unpopular with people in a neighboring country or geographic region.
By the same token, what we tell each other and what others tell us, i.e. Word of Mouth (WoM), shapes our viewpoints and perceptions of the world around us. The English poet John Donne once said: “No man is an island”. Word of mouth is important as it influences how people think and behave, as well as what they buy or do not buy.
In modern society, new communication technologies have expanded the reach of WoM. With the advent of the Internet, WoM now crosses geographic regions, even
continents, reaching a wider audience, and overcoming the limits of time and space. WoM now has a sister: Electronic Word of Mouth (EWoM). It is against this backdrop that the research described in this paper is set.
Consumers appear to trust their relationship with other consumers more than they trust their relationship with brands. In this context, EWoM is an important and convenient source of information for consumers (Cheng & Zhou, 2010). An examination of social network structures reveals that most knots involve communication between peers (Vilpponen, Winter & Sundqvist, 2006). The extent to which people share information within their network is determined by the number of other people in their network who engage in EWoM on any given topic. The more reach shared content develops, the more impact the content has on others in the network (Subramani & Rajagopalan, 2003). Social network structures are based on ties and knots (Vilpponen, Winter & Sundqvist, 2006). The strength of the ties influences message flow and reach (Reingen & Kernan, 1986).
In the context of the social and scientific backdrop described above, the aim of the research described herein was to test the impact of different types of arguments in online reviews on intention to engage in EWoM and to explore the influence of the strength of ties within social networks on intention to engage in EWoM. The research question is as follows:
To what extent does message sidedness in online reviews have an impact on intention to engage in Electronic Word of Mouth, and how does the strength of the tie between the sender and the recipient moderate intention to engage in Electronic Word of Mouth?
2. Theoretical Framework
2.1 From WoM to EWoMWord of Mouth (WoM) is a pervasive social phenomenon that has been around for as long as humans have commanded the power of speech. In the past, stories served as vessels for information and the oral tradition had the broadest reach. Later, print, television and radio usurped part of the role of the oral tradition in the dissemination of information. The advent of the Internet resulted in the development of a new information sharing platforms, including social network sites (Chu & Kim, 2011). Social networks are different from other web-based communication as they include contacts, friends and their opinions, interests and ideas (Chatterjee, 2011). Facebook is the world’s most popular social network site. Just three years after its launch in 2004, Facebook had more than 21 million users (Ellison, Steinfield & Lampe, 2007). Facebook is still growing rapidly, thanks to mobile access to the Internet, which makes it even quicker and easier for users to keep up with their social contacts. Facebook offers users a virtual snapshot of the viewpoints and beliefs of their social contacts at any given moment in time and a fertile ground for Electronic Word of Mouth (EWoM).
2.2 Consumer to Consumer Communication (C2C)
EWoM can be initiated by marketers and the organisations they serve in the form of rewards-based referral mechanisms that encourage EWoM. The EYE Film Museum in Amsterdam, for instance, allows visitors to post a video of themselves superimposed in the silent classic ‘Reis Naar de Maan’ to its GreenScreen site, thus promoting EWoM. Kozinets refers to word of mouth that is not initiated by consumers as amplified word of mouth (Kozinets, Wojnicki, Wilner & Valck, 2010).
Person to person word of mouth, e.g. a review of a purchase from the Wehkamp site, is often referred to as organic Word of Mouth (Kozinets, Wojnicki, Wilner & Valck, 2010), peer to peer communication (Bruyn & Lilien, 2004) or customer to customer communication (Libai, Bolton, Bugel, De Ruyter, Gotz, Risselada & Stephen, 2010).
Communication between consumers is more persuasive than communication between brands and consumers. After all, the hierarchy is flatter between consumers than between brands and consumers: consumers are peers. And peers are a very influential and creditable source of information (Lim, Sia, Lee & Benbasat, 2006). The research described in this paper focuses on communication initiated by and occurring between consumers.
2.3 Impact of EWoM
Recognition of the importance of EWoM by marketers has resulted in the
development of new strategies such as Integrated Marketing Communication, which changes the communication flow from inside->out to outside->in. The predictor for branding strategy is taken from consumer evaluations rather than developed within the organization and then pushed out to the consumer (Kitchen, Brignell & Jones, 2004).
Brands are beginning to see that the most compelling source of information for consumers is other consumers who evaluate their brands and products, and they are very interested in seeing their target group engage in EWoM. EWoM has a positive impact on purchase behavior. EWoM boosts sales and spreads brand and/or product awareness better than any traditional approach. A comparison of EWoM and traditional approaches revealed that EWoM is also a strong factor in the acquisition of new customers. Moreover, EWoM referrals evoke longer and more dynamic responses than advertising (Trusov, Bucklin & Pauwels, 2009).
2.4 Scientific Gap
Engagement in Word of Mouth on communication can be contextualized from different approaches: message characteristics, response types and sender and receiver characteristics (Cheung & Thadani, 2010).Libai, Bolton, Bugel, Ruyter, Gotz, Risselada & Stephen (2010) performed a meta-analysis to study the antecedents and consequences of consumer to consumer interactions. According to the study, the body of available literature investigates the consumers to consumer communication context moderated by channel, product, relational characteristics and market characteristics. The authors found no research testing the impact of message sidedness and tie-strength on intention to engage in EWoM in the setting of social network sites.
Cheng and Zhou (2010) published research results that reveal that message-sidedness is an important factor in determining intention to engage in EWoM, but their research did not explore the impact of tie-strength.
The research described herein attempts to fill this scientific gap.
2.5 Message Sidedness and EWoM
People are becoming more and more resistant to the arguments that they receive. This problem can be overcome by including negative information in the messages. Negative arguments increase message credibility (Crowley & Hoyer, 1994 & Eisen, 2007). Messages that contain positive information and negative information are called two-sided messages. Messages that include only positive arguments are called one-sided messages (Rucker, Petty & Brinol, 2008).
Crowley and Hoyer (1994) tested the persuasion level of two-sided messages. They found that two-sided messages evoke more attention to information. The inclusion of negative information makes content more interesting, thus increasing the reader’s motivation to process the information (incongruity of arguments increase involvement). In addition, two-sided messages enhance brand attitude by increasing source and message credibility. This in turn strengthens intention to purchase. Increased source credibility increases motivation to process the message, increases attention to the message and enhances attitude toward the brand (Eisend, 2010). Greater attention to the message (Crowley and Hoyer, 1994) and higher message credibility (Eisend, 2007) should increase the chance of the message being shared further.
Crowley and Hoyer (1994) and Cheung, Luo, Sia & Chen (2007) prove their findings about the superiority of two-sided messages using explanations based on dual process theories, i.e. Inoculation Theory (McGuire, 1962), Attribution Theory (Kelley, 1967) and Optimal Arousal Theory (After, 1976). Attribution Theory relates to tie-strength and is handled in the section Tie-strength and EWoM. The Inoculation Theory and Optimal Arousal Theory are handled below.
In the Inoculation Theory, the sender includes counterarguments to the advocated position and then proceeds to refute the counterarguments. Furthermore, the sender also warns the receiver about the negative arguments before being exposed to the message. The receiver is more susceptible to positive arguments than to negative arguments refuted post-exposure (Allen, 1991). A study by Hale, Mongeau and Thomas (1991) confirmed the veracity of this theory.
The Optimal Arousal Theory states that motivation to process a message with a moderate amount of negative information will be higher than the motivation to process a
message with no negative information at all. The higher motivation can be explained by the fact that two-sided messages evoke a higher attention level and thus higher involvement (Crowley & Hoyer, 1994; Hale, Mongeau & Thomas, 1991). Eisend agrees, stating that incongruity in the message increases the perceived novelty and thus increases attention to two-sided messages (Eisend, 2007; Eisend 2010).
According to Cacioppo & Petty (1984), two-sided messages are processed through systematic processing. In this type of processing, the receiver evaluates and weighs the positive and negative arguments to arrive at a positive or negative assessment of the overall message. This process increases involvement in the message content, thus making it more likely that the message will be shared further.
Against the backdrop of the research outlined above, i.e. message two-sidedness increases message credibility, attention to the message, involvement in message processing and source credibility, as well as enhancing brand attitude, the first hypothesis of the research is as follows:
H1 Two-sided reviews on Facebook will score higher on intention to engage in EWoM than one-sided reviews on Facebook.
2.6 Tie-strength and EWoM
Cheng and Zhou (2010) isolated four determinants that impact engagement in EWoM in their research: information source, information content, receiver characteristics and the relationship between source and receiver. The receiver evaluates the source based on
attributes that the receiver considers meaningful, e.g. source expertise, source trustworthiness and source credibility (Cheng & Zhou, 2010). The Attribution Theory describes how this
assessment works. In a nutshell, the theory states that consumers attribute causes to effects (Kelley, 1967). The attribution of source characteristics to messages can be seen in
campaigns based on celebrity endorsements. This attribution of the perceived characteristics of the endorser to a product is persuasive (Classical Conditioning Theory - Erdogan, 1999). If the source of a review on Facebook is evaluated as an expert, the source’s expertise is
attributed to the review content.
The assumption for the purposes of the research described in this paper is that when family is the source (i.e. strong social tie), the attribution mechanism will have a stronger effect on the receiver than a friend source (i.e. weak social tie). This assumption can be explained using the concept of Homophily (Cheng & Zhou, 2010) and the Friendship/Liking Principle. Homophily involves the assessment of the extent to which characteristics are shared between the sender and receiver. The Friendship/Liking Principle explains
interpersonal persuasion based on the perceived similarity between communication agents. People tend to like others who seem to be similar to them and dislike others who seem to be divergent to them (Cialdini, 1994). A high level of Homophily can be associated with strong ties and a low level of homophily with weak ties. According to the concept of Homophily, individuals are more apt to find similarities between themselves and family or close friends than between themselves and mere acquaintances. In the context sketched above, the logical assumption is that a message from a strong tie source should evoke more intention to engage in EWoM than a message from weak tie source.
Research performed by Cheung, Luo, & Chen (2007) into the informational influence based determinants of EWoM reveals that source credibility is crucial as it stimulates
engagement in EWoM. The source’s credibility level is determined not only by message characteristics (see Message Sidedness and EWoM in the previous subsection), but also
tie-strength. Chu and Kim (2011) found that tie-strength, trust and informational influence have an impact on WoM behavior.
Traditional WoM is strongly linked to the interaction between family and friends, people with whom we tend to socialize the most and with whom we therefore have strong ties. However, EWoM is more diverse because much of it consists of communication
between communication agents with weak ties. EWoM from a source with whom the receiver has a strong tie relationship with is perceived as more genuine and reliable than EWoM from a sender with whom the receiver has a weak tie. The message receiver is less apt to search for an ulterior motive behind messages exchanged in a sender-receiver relationship with strong ties (Chen & Zhou, 2010). A strong tie evokes higher source credibility, increasing the likelihood that the message will be shared further. Brown & Reingen (1987) tested EWoM in the micro perspective of social networks, and concluded that activation is more likely in cases involving strong and homophilous ties.
Against the backdrop of the research outlined above, i.e. strong tie strength increases source credibility, the second hypothesis of the research is as follows:
H2 Reviews from a sender with whom the receiver has a strong tie will evoke a higher intention to engage in EWoM than reviews from a sender with whom the receiver has a weak tie.
2.7 Message Sidedness, Tie-strength and EWoM
Cheng & Zhou (2010) explain the importance of message and source evaluations in intention to engage in EWoM. Their paper also supports the findings of Cheung, Luo, SIA & Chen (2007). Cheung et al. (2007) tested intention to engage in EWoM in the setting of online consumer review sites, using informational influence, among other factors, as a determinant.
The findings of Cheung et al. (2007) regarding informational influence can be linked to systematic processing of two-sided messages. Both two-sidedness and strong ties increase information credibility (Hou Wee, Luan & Lwin, 1995)
.
Source credibility affects the cognitive response to messages, increasing the likelihood that the message will be seen as a bearer of practical value, making it worthy of sharing further (Berger, 2013). Cheung et al. (2007) states that informational influence is based on the receiver’s perception of content (i.e. message sidedness) and source (i.e. tie-strength).There appears to a relationship between hypothesis H1 and hypothesis H2 of the research described in this paper. This relationship is encapsulated in the third hypothesis of the research:
H3 Two-sided reviews from a strong tie source influence engagement in EWoM more than two-sided reviews from weak tie source and one-sided reviews from a strong tie source influence engagement in EWoM more than two-sided reviews from a weak tie source.
2.8 Motivations of WOM
The research of Henning-Thurau, Gwinner, Walsh and Gremler (2004) tested the following motivations to determine their impact on intention to post a reviews (i.e. engagement in EWoM): motivation driven by a desire to interact socially (i.e. social benefits), motivation driven by concern for others, motivation driven by the opportunity to enhance self-worth (i.e. self-enhancement) and motivation driven by the promise of economic incentives.
Social benefit motivation and self-enhancement motivation are similar concepts to the concept of Social Currency proposed by Berger (2013). According to this concept, the sender sheds a favorable light on himself or herself by sharing his or her opinion. However, in the
case of social benefits, the driver is the sender’s desire to strengthen social links with the receiver of the message. In the case of self-enhancement, the sender attempts to increase his or her self-esteem based on the receipt of favorable responses to the message (Anderson, 1998). Message sidedness or tie-strength can have an impact on the credibility of the message or the credibility of the source. High message credibility and high source credibility increase the intention to engage in EWoM. As mentioned in the subsection Message Sidedness and EWoM, two-sided messages increase the perceived credibility of messages (Inoculation Theory). As mentioned in the subsection Tie-strength and EWoM, strong ties to the message source increase source credibility (Attribution Theory).
Strikingly, the findings of research conducted by Eisingerich, Chun, Liu, Jia & Bell (2014)reveal that most people are less willing to engage in EWoM than in WoM. In fact, Berger (2013) states that only 7% of WoM occurs online. However, people with a relatively high self-enhancement level are more willing to engage in EWoM than in WoM (Eisingerich, Chun, Liu, Jia & Bell, 2014).
In the case of concern for others, the sender is motivated by the assumption that his or her message will be useful to the receiver. The sender’s desire to share his or her positive experience is altruistic (Paul & Paul, 1993; Price, Feick & Guskey, 1995). In organic EWoM, i.e. information sharing initiated by consumers between consumers, high source credibility (i.e. strong-tie source) and high message credibility (i.e. two-sided messages) should result in high levels of motivation driven by concern for others.
According to Verlegh, Ryu, Tuk & Feick, (2013), intention to engage in EWoM is higher in cases involving rewarded EWoM. Message sidedness and tie-strength has an impact on motivation driven by economic benefits (Reingen & Keran, 1986, Wirtz & Chew, 2002).
Against the backdrop of the research outlined above, i.e. message sidedness and tie-strength drive motivations to engage in EWoM, the fourth hypothesis of the research is as follows:
H4 Two-sided messages drive motivations to engage in EWoM more than one -sided messages and strong ties to the review source drive motivations to engage in EWoM more than weak ties.
3. Methods
3.1 Research designThe quantitative experiment had a 2 (one-sided review versus two-sided review) x2 (weak tie with review source versus strong tie with review source) between-subject factorial design. The visualized conceptual model can be found in Appendix V. Despite the lower research validity of a between-subject factorial design in comparison with a within-subject factorial design, this approach was chosen to prevent the participants from guessing the true aim of research. Exposing the individual respondents to both conditions at once would have been a serious threat to the research’s internal validity. As a considerable part of the sample
consisted of graduate students of communication science, the chance of certain participants guessing the aim of the research was quite real.
3.2 Respondents, data collection and procedure
Respondents were recruited through Facebook in order to ensure that the sample was familiar with the Internet and were able to perform the required behavior in an online environment. Much of the recruiting occurred amongst the researcher’s Facebook friends. Moreover, 24 invitations to participate in the research were sent via e-mail to a contact list in the
researcher’s Gmail account. The first question in the questionnaire asked whether the
participant had a Facebook account. If the answer was no, the respondent was excluded from further conduct (n=2). Sampling started with sending private and tailored individual
invitations to participate in the research. In the invitations, the prospective respondents were also asked to share the invitation on their own Facebook page (snowball sampling). In addition, invitations to participate in the research were also posted in the following closed Facebook groups: “ Master Persuasive Communication UvA”, “Master Grad School
was posted in Dutch in the “Respondenten gezocht!” group). The data was gathered using Qualtrics software.
3.3 Conduct and procedure
Respondents were asked to complete a questionnaire that was available via a link in the invitations sent via Facebook, in the invitations posted in the Facebook group and in the invitations sent via Gmail. There were four scenarios for the questionnaire (Appendix VI). Participants were asked to behave as if the scenario were an actual real-life occurrence. Each respondent was assigned to one of four possible scenarios (experimental factors).
Randomization was performed using the randomization tool (randomizer) available in the Qualtrics software. The first experimental factor was the exposure to a one-side framed Facebook review sent from a strong tie source (i.e. mother or father). The second
experimental factor was the exposure to a one-side framed Facebook review sent from a weak tie source (i.e. Facebook friend). The third experimental factor was the exposure to a two-sided framed Facebook review sent from a strong tie source and the fourth experimental factor was the exposure to a two-sided Facebook review sent from a weak tie source. Thus the experiment had 4 between-subject factors (Table 1).
At the beginning of the questionnaire, all participants were given a uniform briefing and informed that the research was being conducted in the context of a thesis under the auspices of the University of Amsterdam. The terms and conditions of the research explained that
participation was ex gratia and that the responses would be handled confidentially.
Prospective respondents were required to accept the terms and conditions of the research in order to proceed to the questionnaire (Appendix VII).
3.4 Stimulus
Exposure in all review scenarios included a referral. The referral consisted of an inactive link to a screenshot of a Facebook brand page with a description of the fictitious brand/product “Just Type”, a writing/translating program. A two-week free trial was offered in the product description (Appendix IV). The need fulfilled by the fictitious brand/product was unspecific in terms of gender, age, education and nationality.
In addition, each respondent was served one of two randomized, fictitious referrals (Facebook review) containing either a one-sided or a two-sided framed message. The opinions in these referrals were deliberately framed in unemotional terms.
Static digitalized print stimulus was used for the study as two-sided messages are processed through the systematic route (Cacioppo & Petty, 1984). In systematic processing, time is required to process the negative arguments in a two-sided review. The extra time required to process the two-sided message derives from the presence of incongruent
arguments present in the message (Boher, Einwiller, Erb & Sieber, 2003).According to the Inoculation Theory, refutational statements in messages increase message persuasiveness in two-sided messages (Allen, 1991, Hale, Mongeau, & Thomas, 1991). The two-sided message therefore included a refutational statement.
3.5 Observed variables and measurement
The construct of the research variables and the level of measurement applied to the data are presented in this paragraph. Most of variables had a 5-point Likert construct, with the exception of demographics, which were operationalized using a categorical level of measurement. All items used to operationalize variables can be found in Appendix VIII. 1. Review sidedness
The independent variable was operationalized through 2 types of review. The first framing (one-sided review) included only positive arguments in the review. The second framing (two-sided review) repeated the positive arguments of the first framing, but also included negative arguments and refutation of these negative arguments.
2. Tie-strength
The intervention was operationalized through 2 conditions of tie-strength: weak social tie and strong social tie. The weak tie source was a Facebook friend. The strong tie source was a parent (i.e. mother or father).
3. EWoM intention
The dependent variable “intention to engage in EWoM’’ was the first measurement after the randomization. This variable measured how likely it was that the message about Just Type would be shared further. The 5-point Likert scale was adopted from the 9-point scale used in a study by Eisngerich et al. (2014) to measure intention to say positive things on Facebook.
4. Motivations on intention
A deeper operationalization of intention to engage in EWoM was achieved through definition of determinants that are motivations on EWoM intention. The determinants were operationalized based on scales developed in the work of Hennig-Thurau,
Gwinner, Walsh & Gremler (2004). 5-point Likert scales, adjusted from the original 7-point scale, were presented with 12 statements to operationalize following
motivations (benefits): a) Social benefits, b) Concern for others benefits, c) Self-enhancement benefits, d) Economic benefits for sender and receiver, e) Economic benefits for sender.
5. Review evaluations
This variable was a manipulation check for the independent variable (i.e. review sidedness). The review evaluation variable had 6 statements with a 5-point Likert scale based on a scale developed by Flanagin & Metzger (2008) to measure message credibility. This manipulation check was applied in the context of research stating that two-sided messages are evaluated as more credible than one-sided messages (Crowley & Hoyer, 1994).
6. Tie-strength evaluations
This variable had 4 statements with a 5-point Likert scale. The scale was based on a 10-point Likert scale developed by Ha (1996) to rate the closeness to and intimacy with a source and to rate the likelihood of sharing personal information and the likelihood of engaging in activities with the source. The variable was a manipulation check to conditions of intervention, i.e. tie-strength.
4. Results
4.1
Pre-analyses
4.1.1 Demographics
A general picture of the sample composition (N= 201) was developed by measuring the following characteristics: age category, gender, level of education and nationality (Appendix I). Age was measured using 5 age categories. 86% of the respondents in the sample were individuals between 18 and 34 years. The sample consisted of 38% males (n= 77) and 68% females (n=124). Substantial part of the sample (90%) consisted of people with a higher education. 43% of the respondents had a Bachelor’s degrees and 57% of the respondents had a Master’s degrees or higher. Therefore the sample distribution was abnormal. The
distribution of nationality was bimodal. This variable was therefore recoded at binary level to arrive 49% Dutch and 51% of other nationalities. More demographic details can be found in Appendix I.
4.1.2 Randomization check
Chi-square analyses were conducted to determine whether conditions of the independent variable (i.e. message sidedness) differed on respondent characteristics, which were measured using a categorical level of measurement.
Age 2(4)=0.96, p=.915, Gender 2(1)=0.22, p=.636, Education level 2(4)=1.70, p=.793 and Nationality (Dutch vs Others) 2(1)=2.44, p=.785 did not differ significantly between
conditions of review type (i.e. message sidedness). Chi-square analyses were also conducted to check the differences between conditions of the moderator variable (i.e. tie-strength). Age
2
(4)=0.93, p=.920, Gender 2(1)=1.30, p=.250, Education level 2(4)=1.09, p=.896 and Nationality (Dutch vs Others) 2(1)=3.61, p=.060 did not differ significantly between conditions of the review type (i.e. message sidedness). Descriptive statistics for the
manipulation of the independent variable and for intervention conditions can be found in Appendix III. The randomization check showed no significant differences between factorial conditions with regard to age, gender, education and nationality, which means that the null hypothesis was supported.
4.1.3 Reliability analyses
Tests of reliability showed that the scales had internal consistency ranging from good ( ≥ 0.7) to excellent ( ≥ 0.9). The exception of an unacceptable internal consistency occurred in the Product Evaluation measurement under the factor two-sided review and weak source tie ( = .44). Removing the item that evaluated product as “reminding of another product” improved internal consistency to an acceptable level (=.66). All reliability scale results can be found in Appendix II.
4.1.4 Manipulation checks
The manipulation check for review type (i.e. message sidedness) was a 5-point variable measurement of review evaluations. The manipulation check confirmed manipulation as respondents exposed tothe two-sided review evaluated the review type higher (M= 3.42, SD= 1.245) than respondents exposed to the one-sided review (M= 2.97, SD= 1.072). This
difference was significant t(199)= -2.736 p<.05, 95% CI [-0.77, -0.12]. In order to check the difference of intervention conditions (i.e. weak tie versus strong tie), a 5-point variable measurement was conducted on tie-strength of the review source. Evaluations of tie-strength were higher for respondents who were exposed to a review from a strong tie source (M=3.50, SD= 1.22) than for respondents who were exposed to a review from a weak tie source
(M=3.07, SD= 1.17). The difference was statistically significant t(199)= -2.60 p<.05, 95% CI [-0.77, -0.10]. Significant differences of message sidedness on review evaluations and
operationalized, i.e. that the variables measured what they were supposed to be measuring (internal validity).
4.2
Testing hypotheses
4.2.1 Main effect
Based on the fact that the experiment had a between-subject design, Hypothesis 1 and Hypothesis 2 were tested through independent t-test analyses. Respondents exposed to a one-sided review had less intention to engage in EWoM than respondents exposed to a two-one-sided review. Hypothesis 1 was not supported since the difference between one-sided and two-sided reviews on intention to engage in EWoM was not statistically significant p>.05 (Table 2)
Table 2 Results of t-test and Descriptive Statistics for intention to engage in EWoM by sidedness of review
Review 95% CI for Mean
Difference One-sided Two-sided M SD n M SD n t df Intention to engage in EWoM. 2.54 1.44 108 2.71 1.60 93 -0.60 0.25 0.80 199
The independent t-test also showed results converse to Hypothesis 2. Respondents exposed to a review from a week tie source had more intention to engage in EWoM than respondents exposed to a review from a strong tie source. This difference was not statistically significant either, therefore Hypothesis 2 was also rejected (Table 3)
Table 3 Results of t-test and Descriptive Statistics for intention to engage in EWoM by sidedness by tie-strength
Review 95% CI for Mean
Difference Weak tie Strong tie
M SD n M SD n t df Intention to engage in EWoM. 2.73 1.56 102 2.51 1.47 99 -0.20 0.64 0.58 199 4.2.2 Interaction
Hypothesis H2 tested the same effect of sidedness on intention to engage in EWoM as
Hypothesis H1, but with the moderation of tie-strength (tested in Hypothesis H2). Intention to engage in EWoM scores was subjected to a two-way analysis of variance, with two types of review (one-sided, two-sided) and two levels of tie-strength (high, low) (Table 4).
Confirming the results from the independent t-tests of Hypothesis 1 and Hypothesis 2, there was a non-significant main effect of review sidedness on intention to engage in EWoM F(1,197)= 0,54 p=.461 and there was a non-significant effect of tie-strength of review source on intention to engage in EWoM F(1,197)= 0,93 p=.336. The interaction effect was not statistically significant F(1,197)= 0,006 p=.938 (Figure 1).
Table 4 Descriptive Statistics of review sidedness and tie-strength of review source on intention to engage in EWoM
Source tie-strength
Review sidedness Mean Std. Deviation N
Weak tie-strength of review source 1 sided 2.64 1.425 50 2 sided 2.82 1.704 49 Total 2.73 1.564 99 Strong tie-strength of review source 1 sided 2.45 1.465 58 2 sided 2.59 1.499 44 Total 2.51 1.474 102 Total 1 sided 2.54 1.443 108 2 sided 2.71 1.606 93 Total 2.62 1.519 201
Figure 1 Lack of interaction effect between review sidedness and tie-strength of review source on intention to engage in EWoM.
4.2.3 EWoM Motivations and Sidedness
Independent t-test analyses were conducted for each motivation. The t-test findings showed a non-significant difference t(199)= 0.52 p>.05, 95% CI [-0.23, 0.41] between respondents exposed to a one-sided review (M=2.96, SD= 1.06) and respondents exposed to a two-sided review (M=2.87, SD= 1.23) on motivations driven by social benefits. A non-significant difference t(199)= -0.169 p>.05, 95% CI [-0.34, 0.29] was found between the effect of a one-sided review (M=3.50, SD= 1.05) and a two-one-sided review (M=3.53, SD= 1.19) on motivation driven by the determinant concern for others. Although respondents exposed to a one-sided review (M=2.69, SD= 1.03) scored higher on motivation driven by self-enhancement than respondents exposed to a two-sided review (M=2.62, SD= 1.15), the difference was not statistically significant t(199)= 0.40 p>.05, 95% CI [-0.24, 0.36]. There was a non-significant difference t(199)= -0.37 p>.05, 95% CI [-0.45, 0.31] between the effect of one-sided framing (M=2.44, SD= 1.31) and two-sided framing (M=2.52, SD= 1.41) on motivation driven by economic benefits for the sender. Lastly, the one-sided review (M=2.75, SD= 1.43) evoked less motivation driven by economic benefits for the sender and the receiver than the twosided review (M=2.88, SD= 1.55). This difference was also statistically insignificant t(199)= -0.63 p>.05, 95% CI [-0.55, 0.28].
In summary, one-sided messages scored higher than two-sided messages on motivations driven by social benefits and self-enhancement. Two-sided messages scored higher than one-sided messages on motivations driven by concern for others, economic benefits for the sender and economic benefits for the sender and receiver.
4.2.4 EWoM Motivations and Tie-strength
Tie-strength was also tested on motivations to engage in EWoM. There was a non-significant difference t(199)= 1.10 p>.05, (95% CI [-0.14, 0.50] between a strong tie with the source of
the review (M=2.83, SD= 1.17) and a weak tie with the source of the review (M=3.01, SD= 1.11) in the score on evaluation of social benefits as a determinant of intention to engage in EWoM. There was also a non-significant difference t(199)= 1.04 p>.05, 95% CI [-0.15, 0.48] between respondents who saw the review from a strong tie (M=3.43, SD= 1.13) and
respondents who saw the review from a weak tie (M=3.60, SD= 1.11) in the score on
evaluation of concern for others as a determinant of intention to Engage in EWoM. There was a non-significant difference t(199)= 0.26 p>.05, 95% CI [-0.26, 0.34] between strong tie source (M=2.64, SD= 1.10) and weak tie source (M=2.68, SD= 1.08) in the score on motivation driven by self-enhancement. Motivation driven by economic benefits for the sender did not differ t(199)= -0.76 p>.05, 95% CI [-0.52, 0.23] between the review from strong tie source (M=2.55, SD= 1.32) and the review from a weak tie source (M=2.40, SD= 1.40). Although the review from a strong tie source (M=2.77, SD= 1.47) had higher scores on motivation driven by economic benefits for the sender and receiver than the review from a weak tie source (M=2.85, SD= 1.50), the difference was not statistically significant t(199)= 0.35 p>.05, 95% CI [-0.34, 0.49].
In summary, the review from a strong tie source scored higher than the review from a weak tie source on motivation driven by economic benefits for the sender. The review from a weak tie source scored higher than the review from a strong tie source on motivation driven by social benefits, concern for others, self-enhancement and economic benefits for the sender and receiver. This result contradicts the assumption that strong tie-strength will result in higher scores on motivations to engage in EWoM than weak strength.
In conclusion, Hypothesis H4 was not supported by the findings. The analyses of the effect of message-sidedness and tie-strength on motivations to engage in EWoM yielded statistically insignificant results.
4.3 Covariates
In order to test the construct the conceptual model, demographic variables were tested as covariates. Age F(1,195)=0.699, p=.404, gender F(1, 195)=0.412, p=.521, education F(1, 195)=1.25, p=.265 and nationality F(1, 195)=1.157, p=.283 had non-significant effects on the construct of the conceptual model.
5. Conclusion and Discussion
Given that the findings of the research do not support or confirm the hypotheses, the research question remains unanswered.
Hypotheses H1 and H2, which related to message sidedness and tie-strength
respectively, were not supported by the findings of this research. Although the findings did show a difference, the difference was not statistically significant. Hypothesis H3 related to the interaction between message sidedness and tie-strength and their shared impact on intention to engage in EWoM and was in fact a combination of Hypotheses H1 and H2. Given the fact that the findings for Hypotheses H1 and H2 were statistically non-significant, it is no surprise that the findings do not support Hypothesis H3. Hypothesis H4, which related to message sidedness and tie-strength as drivers for motivations of intention to engage in EWoM, was also not confirmed by the findings.
It is worth noting that the research was constrained by a number of limitations.
Firstly, the size of the sample was smaller than desired, which means that the research suffers from reduced research reliability. A larger sample size would have increased research reliability. According to Reingen & Kernan (1986), consumer networks should be tested on a large scale because EWoM is a social phenomeno n.
Secondly, the distribution of sample demographics in the research was abnormal, particularly in terms of age and education. Most of the participants were in their mid-twenties, with a higher education. The number of participants in other age groups and with other educational levels was extremely low.
Thirdly, the setting of the research was artificial and the behavior being tested was hypothetical. The participants did not actually receive a review written by a parent or friend. The review was written by the researcher and served to them in the questionnaire. The review
was a review of an imaginary brand. No real economic benefits applied thus motivations driven by economic benefits did not apply. The other motivations to engage in EWoM were measured based on the participant’s ability to imagine him or herself in a purely hypothetical situation.
Fourthly, in an attempt to overcome the limitation of an artificial setting and hypothetical behavior, the participants were asked to use their imaginations. Not everyone has an equal amount of imagination. The hypothetical setting required an imitation of behavior as if it was performed in reality, which required adherence to a fairly complicated set of instructions. Many of the participants were gathered through Facebook posts on closed student groups. Users of these closed student groups were primarily students writing theses for their master’s degree, which implies a certain amount of time pressure, which in turn may have resulted in a lack of adherence to the instructions. This may have diminished the validity of the research.
The purpose of this study was to add to the existing body of knowledge on the topic of intention to engage in EWoM. Although, as stated above, the research did not support or confirm the hypotheses or answer the research question, the research does have implications for future research. For instance, a follow-up study that eliminates the limitations described above could enrich the body of knowledge on the topic of EWoM.
Once such a follow-up study is completed and questions related to the interaction of message sidedness and tie-strength are satisfactorily answered, additional determinants such as product compatibility with consumer need, prior knowledge of the product (Lim, Sia,Lee & Benbasat, 2006; Libai, Bolton, Bugel, Gotz, Risselada & Stephen, 2010) and visual cues relating to the product (Davis & Khazanchi, 2008) could be tested. All these product characteristics influence EWoM behavior and intention to engage in EWoM.
Another determinant to intention to engage in EWoM is consumer skepticism to attributes of products and messages about products (Lee and Youn, 2009). It would also be worthwhile to test this determinant.
All in all, the use of message sidedness and tie-strength as determinants on intention to engage in EWoM in the setting of Facebook was a novel and unique approach in the body of literature on the topic of message sidedness (Allen, 1991; Crowley & Hoyer, 1994; Hale et at, 1911; Hou Wee et al., 1995; Eisend, 2007 & 2010; Verlegh et al., 2013 ) and the topic of tie-strength ( Reingen & Kernan, 1986; Browm & Reingen, 1987; Wirtz & Chew, 2002; De Bruyn & Lilien, 2004; Vilpponen, Winter & Sundqvist, 2006; Cheng & Zhou, 2010; Chu & Kim, 2011) on EWoM.
In a more social context, a deeper understanding of the impact of message sidedness and tie-strength on intention to engage in EWoM would help brands optimize strategy
development and strategy execution. According to the Theory of Planned Behavior proposed by Ajzen (2011), intention is a main determinant to behavior execution. Hence, a greater understanding of the determinants of consumer intention to engage in EWoM implies a deeper and richer understanding of the wider science of EWoM behavior.
The findings of this research in the consumer-to-consumer (C2C) setting of EWoM can be transferred to the strategies used in the dimension of EWoM initiated by brands (Kitchen, Brignell & Jones, 2004). This transfer can be particularly valuable for brands that are new on the market. These brands cannot afford the expense of traditional advertising. EWoM is also a relatively cheap way to acquire customers and it produces higher dynamic effects than traditional advertising (Trusov, Bucklin & Pauwels, 2009). In addition to a larger reach than traditional media, social network sites also allow targeted messaging. New brands and
research in this paper to develop viral campaigns that operate in the setting of EWoM. (E)WoM is a powerful tool to shape opinions regarding products and brand image.
References
AFTER, M. J. (1976). Some data inconsistent with the optimal arousal theory of motivation. Perceptual and Motor Skills, 43(3f), 1209-1210.
Ajzen, I. (2011). Theory of planned behavior. Handb Theor Soc Psychol Vol One, 1, 438. Allen, M. (1991). Meta‐ analysis comparing the persuasiveness of one-sided and
two-sided messages. Western Journal of Speech Communication, 55(4), 390-404. Anderson, E. W. (1998). Customer satisfaction and word of mouth. Journal of service
research, 1(1), 5-17.
Berger, J. (2013). Contagious: Why things catch on. Simon and Schuster.
Brown, J. J., & Reingen, P. H. (1987). Social ties and word-of-mouth referral behavior.
Journal of Consumer research, 350-362.
Cacioppo, J. T., & Petty, R. E. (1984). The elaboration likelihood model of persuasion.
Advances in consumer research, 11(1), 673-675.
Chatterjee, P. (2011). Drivers of new product recommending and referral behavior on social network sites. International Journal of Advertising, 30(1), 77-101. Cheng, X., & Zhou, M. (2010, August). Study on effect of eWOM: A literature review
and suggestions for future research. In Management and Service Science
(MASS), 2010 International Conference on (pp. 1-4). IEEE.
Cheung, C. M., & Thadani, D. R. (2010). The effectiveness of electronic word-of-mouth communication: A literature analysis. Proceedings of the 23rd Bled
eConference eTrust: Implications for the Individual, Enterprises and Society,
329-345.
Cheung, M. Y., Luo, C., SIA, C. L., & Chen, H. (2007). How do people evaluate
electronic word-of-mouth? Informational and normative based determinants of perceived credibility of online consumer recommendations in China.
PACIS 2007 Proceedings, 18.
Cialdini, R. B. (1994). Interpersonal influence.
Chu, S. C., & Kim, Y. (2011). Determinants of consumer engagement in electronic word-of-mouth (eWOM) in social networking sites. International journal of
advertising, 30(1), 47-75.
Crowley, A. E., & Hoyer, W. D. (1994). An integrative framework for understanding two-sided persuasion. Journal of Consumer research, 561-574.
Davis, A., & Khazanchi, D. (2008). An empirical study of online word of mouth as a predictor for multi-product category e-commerce sales. Electronic Markets, 18(2), 130-141.
De Bruyn, A., & Lilien, G. L. (2008). A multi-stage model of word-of-mouth influence through viral marketing. International Journal of Research in Marketing, 25(3), 151-16
Eisend, M. (2007). Understanding two‐ sided persuasion: An empirical assessment of theoretical approaches. Psychology & Marketing, 24(7), 615-640.
Eisend, M. (2010). Explaining the joint effect of source credibility and negativity of information in two-sided messages. Psychology & Marketing, 27(11), 1032-1049.
Eisingerich, A. B., Chun, H. H., Liu, Y., Jia, H. M., & Bell, S. J. (2015). Why
recommend a brand face-to-face but not on Facebook? How word-of-mouth on online social sites differs from traditional word-of-mouth. J Consum
Psychol, 25, 120-128.
Ellison, N. B., Steinfield, C., & Lampe, C. (2007). The benefits of Facebook “friends:” Social capital and college students’ use of online social network sites. Journal
of Computer‐ Mediated Communication, 12(4), 1143-1168.
Erdogan, B. Z. (1999). Celebrity endorsement: A literature review. Journal of marketing
management, 15(4), 291-314.
Flanagin, A. J., & Metzger, M. J. (2008). The credibility of volunteered geographic information. GeoJournal, 72(3-4), 137-148.
Ha, L. (1996). Observations: advertising clutter in consumer magazines: dimensions and effects. Journal of Advertising Research, 36, 76-84.
Hale, J. L., Mongeau, P. A., & Thomas, R. M. (1991). Cognitive processing of one‐ and two‐ sided persuasive messages. Western Journal of Speech Communication, 55(4), 380-389.
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.
Hou Wee, C., Luan Lim, S., & Lwin, M. (1995). Word-of-mouth communication in Singapore: With focus on effects of message -sidedness, source and user-type.
Asia Pacific Journal of Marketing and Logistics, 7(1/2), 5-36.
Kelley, H. H. (1967). Attribution theory in social psychology. In Nebraska symposium on
Kitchen, P. J., Brignell, J., Li, T., & Jones, G. S. (2004). The emergence of IMC: a theoretical perspective. Journal of advertising research, 44(01), 19-30.
Kozinets, R., De Valck, K., Wojnicki, A., & Wilner, S. (2010). Networked Narratives: Understanding Word-of-Mouth Marketing in Online Communities. Journal
Of Marketing, 74(2), 71-89.
Lee, M., & Youn, S. (2009). Electronic word of mouth (eWOM) How eWOM platforms influence consumer product judgement. International Journal of Advertising, 28(3), 473-499.
Libai, B., Bolton, R., Bügel, M. S., De Ruyter, K., Götz, O., Risselada, H., & Stephen, A. T. (2010). Customer-to-customer interactions: broadening the scope of word of mouth research. Journal of Service Research, 13(3), 267-282.
Lim, K. H., Sia, C. L., Lee, M. K., & Benbasat, I. (2006). How do I trust you online, and if so, will I buy? An empirical study on designing Web contents to develop online trust. Journal of Management Information Systems, (23), 233-266. McGuire, W. J., & Papageorgis, D. (1962). Effectiveness of forewarning in developing
resistance to persuasion. Public Opinion Quarterly, 26(1), 24-34. Paul, E. F., & Paul, J. (Eds.). (1993). Altruism. Cambridge University Press. Price, L. L., Feick, L. F., & Guskey, A. (1995). Everyday market helping behavior.
Journal of Public Policy & Marketing, 255-266.
Reingen, P. H., & Kernan, J. B. (1986). Analysis of referral networks in marketing: Methods and illustration. Journal of Marketing Research, 370-378. Rucker, D. D., Petty, R. E., & Briñol, P. (2008). What's in a frame anyway?: A
meta-cognitive analysis of the impact of one versus two sided message framing on attitude certainty. Journal of Consumer Psychology, 18(2), 137-149.
Subramani, M. R., & Rajagopalan, B. (2003). Knowledge-sharing and influence in online social networks via viral marketing. Communications of the ACM, 46(12), 300-307.
Trusov, M., Bucklin, R. E., & Pauwels, K. (2009). Effects of word-of-mouth versus traditional marketing: Findings from an internet social networking site.
Journal of marketing, 73(5), 90-102.
Verlegh, P. W., Ryu, G., Tuk, M. A., & Feick, L. (2013). Receiver responses to rewarded referrals: the motive inferences framework. Journal of the Academy of
Marketing Science, 41(6), 669-682
Vilpponen, A., Winter, S., & Sundqvist, S. (2006). Electronic word-of-mouth in online environments: Exploring referral networks structure and adoption behavior.
Wirtz, J., & Chew, P. (2002). The effects of incentives, deal proneness, satisfaction and tie strength on word-of-mouth behavior. International journal of service
APPENDIX III
Descriptive statistics for main manipulation and intervention
conditions
APPENDIX IV
Stimuli
Screenshot of Facebook brand page
One-sided review
Just Type is an easy way to communicate with people in other countries. It makes perfect translations!
Two-sided review
Translations in Just Type are not always completely correct in terms of structure and grammar but the errors it makes are minor. Just Type is an easy way to communicate with people in other countries.
APPENDIX V
Conceptual model
APPENDIX VI
Scenarios
Scenario 1: one-sided message and strong tie source
Imagine that you got a notification on your Facebook account that your father or mother posted on his or her Facebook wall following review:
Just Type is an easy way to communicate with people in other countries. It makes perfect translations! Please remember this review and remember that the review is posted by your mother or father. Please answer following questions keeping that in your mind.
Scenario 2: one-sided message and weak tie source
Imagine that you got a notification on your Facebook account that your friend posted on his/her Facebook wall following review:
Just Type is an easy way to communicate with people in other countries. It makes perfect translations! Please remember this review and remember that the review is posted by your specific Facebook friend. Please imagine your real Facebook friend and answer following questions keeping that in your mind.
Scenario 3: two-sided message and strong tie source
Imagine that you got a notification on your Facebook account that your father or mother posted on his or her Facebook wall a following review:
Translations in Just Type are not always completely correct in terms of structure and grammar but the errors it makes are minor. Just Type is an easy way to communicate with people in other countries.
Please remember this review and remember that the review is posted by your mother or father. Please answer following questions keeping that in your mind.
Scenario 4: two-sided message and weak tie source
Imagine that you got a notification on your Facebook account that your friend posted on his/her Facebook wall following review:
Translations in Just Type are not always completely correct in terms of structure and grammar but the errors it makes are minor. Just Type is an easy way to communicate with people in other countries! Please remember this review and remember that the review is posted by your specific Facebook fr iend. Please imagine your real Facebook friend and answer following questions keeping that in your mind.
APPENDIX VII
Invitation to participate in research and research terms &
conditions
Dear Sir / Madam,
I would like to invite you to participate in a study to be conducted under the auspices of the Graduate School of Communication, a part of the University of Amsterdam.
During the study, you will be asked to assess the review about fictitious brand: JUST TYPE. At the end of the survey, you will be asked to answer a number of demographical questions. The goal of this research is to analyse information sharing behaviour. The study will take about only 5 minutes of your time.
As this research is being carried out under the responsibility of the ASCoR, University of Amsterdam, we can guarantee that:
1) Your anonymity will be safeguarded, and that your personal information will not be passed on to third parties under any conditions, unless you first give your express permission for this.
2) You can refuse to participate in the research or cut short your participation without having to give a reason for doing so. You also have up to 24 hours after participating to withdraw your permission to allow your answers or data to be used in the research.
3) Participating in the research will not entail you being subjected to any appreciable risk or discomfort, the researchers will not deliberately mislead you, and you will not be exposed to any explicitly offensive material.
If you are under age of 18, please do not participate in this study.
For more information about the research and the invitation to participate, you are welcome to contact Tomasz Kawecki (tomas.kawecki27@gmail.com) at any time.
Should you have any complaints or comments about the course of the research and the procedures it involves as a consequence of your participation in this research, you can contact the designated member of the Ethics Committee representing ASCoR, at the following address: ASCoR Secretariat, Ethics Committee, University of Amsterdam, Postbus 15793, 1001 NG Amsterdam; 020‐525 3680; ascor‐secr‐fmg@uva.nl.
Any complaints or comments will be treated in the strictest confidence.
We hope that we have provided you with sufficient information. We would like to take this opportunity to thank you in advance for your assistance with this research, which we greatly appreciate.
Kind regards, Tomasz Kawecki
