M A N I P U L A T I O N O F T H E O N L I N E
R E V I E W S : A P R O F I T A B L E
M A R K E T I N G A P P R O A C H O R
C O U N T E R P R O D U C T I V E ?
WHAT IS THE ROLE OF BRAND KNOWLEDGE ON CONSUMER’S PURCHASE
INTENTIONS WITH THE KNOWLEDGE ABOUT MANIPULATED ONLINE REVIEWS?
Student
Edwin Kurvers
Student number
5818273
Course
Master thesis
Institution
Amsterdam Business School / University of Amsterdam
Study
MSc Business Administration – Marketing track
Supervisor
Dr. F. Situmeang
Date
19-1-2016
Statement of Originality
This document is written by Edwin Kurvers who declares to take full responsibility for the
contents of this document.
I declare that the text and the work presented in this document is original and that no sources
other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of
completion of the work, not for the contents.
Acknowledgement
I have reached the final step in completing the study Business Administration at the
University of Amsterdam, with the completion of this thesis. The track of the study is
marketing and therefore I knew from the beginning of choosing a subject that online reviews
is a very interesting topic to research.
Firstly, I would like to take the opportunity to thank my supervisor dr. Frederik Situmeang
for all of his positive support and guidance during the process of writing this thesis.
Furthermore, I would like to thank my family and friends who have supported me and
especially my girlfriend Tessa Duin.
I hope you enjoy reading my thesis.
Kind regards,
Abstract
Consumers are more conscious about the marketing approaches that organizations
manipulate online reviews to increase their sales. However, consumers increasingly rely on
online reviews, because the product quality is difficult to evaluate before consumption.
Therefore, this study will investigate the relationship between the knowledge about
manipulation of the online reviews and the purchase intentions. Thereby, we will investigate
the role of brand knowledge in this relationship with two drivers: brand awareness and brand
image. We used the following research question in this study: “What is the role of brand
knowledge on consumer’s purchase intentions with the knowledge about manipulated online
reviews?”
Using an experimental survey data from 592 respondents, we found that there is a
significant positive relationship between the knowledge about manipulation of the online
reviews and the purchase intentions. Moreover, we found that both brand awareness and
brand image have a significant moderating effect on this positive relationship, so that the
relationship is stronger for lower values of brand awareness and brand image than for higher
values of brand awareness and brand image. Further, we discuss the results of this study,
implications, limitations and do suggestions for future research.
TABLE OF C ONTENTS
1. Introduction ... 7 1.1 Literature Gap ... 9 1.2 Scientific relevance ... 9 1.3 Managerial relevance ... 10 2. Literature review ... 11 2.1 Online reviews ... 11 2.2 Purchase intentions ... 122.2.1 Relationship between online reviews and purchase intentions ... 12
2.3 Manipulated online reviews ... 13
2.3.1 Relationship between manipulated online reviews and purchase intentions ... 15
2.4 Brand knowledge ... 16
2.3.1 Brand awareness ... 16
2.3.2 Moderating role of brand awareness ... 17
2.3.3 Brand image ... 18
2.3.4 Moderating role of brand image ... 19
3. Conceptual model ... 21 3.1 Hypothesis... 21 4. Research design ... 22 4.1 Sample... 22 4.2 Research design ... 23 4.2.1 Survey procedure ... 24 4.2.2 Measures ... 26
4.3 Procedure data collection ... 27
4.3.1 Pilot studies ... 27
4.3.2 Main study ... 27
5. Results and analysis ... 29
5.1 Preliminary study ... 29
5.3 Analysis results ... 31
6. Conclusion and discussion ... 38
7. Implications, limitations and future research ... 42
8. References ... 44
Appendix 1: Data results pretest ... 51
1.
INTRODUCTION
Today, online reviews is an important topic in the marketing science. Online reviews can be defined as peer generated or user-generated product evaluations (Jalilvald, Esfahani & Samiei, 2011). Jimenez and Mendoza (2013) describe online reviews as having graphical and/or textual elements that allow consumers to assess and determine the usefulness of the reviews for their purchase decision. Before purchasing an item, 81 percent of the customers read the reviews (Deloitte, 2014) and 82 percent of the consumers say their purchase decision have been directly influenced by the product reviews (Deloitte, 2007). Zhang et al. (2014) argue that online reviews affects consumers’ purchase decision. Often by e-commerce reliable information about the product is hard to find, because the product quality is difficult to evaluate before consumption (Situmeang, Leenders & Wijnberg, 2014). That is why consumers use the online reviews as a signal of the quality of the product (Situmeang, Leenders & Wijnberg, 2014) and aid the online reviews in their purchase decisions (Nielsen, 2012). Several studies investigate how online reviews influence consumers’ purchase intentions (Park et al., 2007; Zhang et al., 2007; Zhang et al., 2014; Lee, 2009; Bae & Lee, 2011; Zou et al., 2011). However, all these studies assume that the online reviews are honest reviews submitted by real consumers.
As marketing approach, companies can use the online reviews by manipulation to increase their sales. Following previous literature, reviews manipulation can be defined as “vendors, publishers, writers or any third-party consistently monitoring, the online reviews and posting
non-authentic online reviews on behalf of customers when needed, with the goal of boasting the sales of
their products” (Hu, Liu and Sambamurthy, 2011). In their study with hotel reviews, Hu et al. (2012)
find that 10.3% of the products are subject to online reviews manipulation by examining the writing style. They also argue that consumers are only able to detect the manipulation taking place through ratings, but not through sentiments. Another example of manipulation is the software manipulation of the diesel cars by Volkswagen. Thereby, Volkswagen has sold millions of cars to consumers who had confidence in the reliability of the software, while the American environmental inspections turned out
to have been manipulated1. In this case, there is not a specific manipulation of reviews. However, the manipulation will have an impact on consumer purchase intentions, because this scandal can lead to a breach of trust with consumers. Overall, it is still difficult to find out whether reviews for products or services are manipulated. Review manipulation is a new phenomenon and that is why there is limited understanding about how to detect review manipulation (Hu, Bose, Gao and Liu, 2011). Nevertheless, Bambauer-Sachse and Mangold (2013) argue that consumers are more conscious about the marketing approaches of manipulating online reviews. To the best of my knowledge, no research has been done about manipulation of online reviews and their impacts on the purchase intentions of consumers.
The present research will study the relationship between the impact of knowledge about manipulated online reviews on the consumers purchase intentions and the role of the brand knowledge. From the example about the Volkswagen scandal, the knowledge of manipulation by consumers has an impact on the consumer relationship with the brand and overall this breach of trust can lead to an adjustment of the purchase intentions. Thereby, it is very interesting to understand the role of brand knowledge in the situation that when consumers know the online reviews are manipulated. Over the last decades, organizations have increased the investments in the development of their customer- based brand equity (Keller, 2003; Aaker and Biel, 1992). Keller (1993) defines “Customer-based brand equity as the differential effect of brand knowledge on consumer response to the marketing of the brand”. Peter and Olson (2001) state that consumer brand knowledge relates to
the cognitive representation of the brand. Moreover, Keller (1993) argue that “Brand knowledge is conceptualized according to an associative network memory model in terms of two components;
brand awareness and brand image”. Several studies investigate the role of brand knowledge and the effects on consumers purchase intentions (Jalilvand and Samiei, 2012; Wu et al., 2011; Bambauer-Sachse and Mangold, 2011; Wang and Yang, 2010). However, the current research will investigate the role of brand knowledge on the relationship between the knowledge about manipulated online reviews and the purchase intentions of consumers.
1.1 LITERATURE GAP
Current literature have been conducted to examine the relationship of online reviews and consumers’ purchase intentions. Even the effects of the characteristics of online reviews on the consumers’ intentions to buy are researched. Little research has been done about the manipulation of online reviews, because it is hard to know whether a review is fake or general. However, to the best of my knowledge, none research has been done about the role of brand knowledge in combination with the knowledge about manipulation of the online reviews. So overall, this study is aimed at closing the gap in the existing literature about the role of brand knowledge. The main research question is established as follows: What is the role of brand knowledge on the negative relationship between the knowledge about manipulation of the online reviews and consumers’ purchase intentions?
1.2 SCIENTIFIC RELEVANCE
From a scientific perspective research have been done about the effects of the knowledge about manipulated online reviews. The impact of manipulated online reviews is difficult to measure because the consumers could not yet know that a review actually is manipulated before consumption. Of course consumers could have the suspicion that a review is manipulated. However, scientists have started investigating the phenomenon of manipulation of the online reviews. That is the reason that consumers are more conscious about this marketing approach. In such a situation there is the need which factors influence the purchase decision of consumers. Previous studies have done research about the internal factors of online reviews or about the process of online reviews on the purchase decision of customers. There have not been done many research to previous experiences with the brand. That is why there is the need to know whether the role of brand knowledge influence the purchase decision of consumers by online reviews with the knowledge that reviews could be manipulated.
1.3 MANAGERIAL RELEVANCE
From a managerial perspective, online reviews have a big impact on the purchase decisions of consumers. Consumers are more skeptical about the manipulation of the review, but the knowledge that reviews can be manipulated does not have an impact on the product evaluations when only positive reviews are encountered (Bambauer-Sachse & Mangold, 2013). However, by a lack of consensus the detrimental effect of the negative reviews on the positive reviews could have an impact on the purchase decision of the consumer. Previous studies are only focused on the credibility of the source of the reviews (internal factor), but not on the knowledge about the brand (external factor). For managers, it is interested to know whether the brand knowledge influence the product evaluations of consumers with the understanding that the online reviews are manipulated.
2. LITERATURE REVIEW
This chapter provides a comprehensive review of the literature about the key concepts in this study. In order to analyze what has already been researched about brand knowledge and brand awareness on purchase intentions of consumers. First some basic information about online reviews and their impact on the purchase intention is given. We explore why we use the term intention instead of decision. Next, some examples of manipulated reviews are given and how consumers react on the knowledge about these manipulations. Previous research about the impact on consumers’ buying intentions is given. Further, previous research about brand knowledge and brand awareness is given. Also the impact of both on the purchase intentions of consumers and the impact of manipulated reviews on the brand knowledge and brand awareness are explored. Then, a detailed analysis of the possible relations between the role of the brand and manipulated reviews on the purchase intentions are defined. Finally, based on the review, a conceptual framework and hypotheses are developed.
2.1 ONLINE REVIEWS
Online reviews can be defined as peer generated or user-generated product evaluations (Jalilvald, Esfahani & Samiei, 2011). Jimenez and Mendoza (2013) describe online reviews as having graphical and/or textual elements that allow consumers to assess and determine the usefulness of the reviews for their purchase decision. Most of all, the online reviews involve direct or indirect recommendations (Willemsen, 2013). Before purchasing an item, 81 percent of the customers read the reviews (Deloitte, 2014) and 82 percent of the consumers say their purchase decision have been directly influenced by the product reviews (Deloitte, 2007). Zhang et al. (2014) argue that online affects consumers’ purchase decision. Often by e-commerce reliable information about the product is hard to find, because the product quality is difficult to evaluate before consumption (Situmeang, Leenders & Wijnberg, 2014). That is why consumers use the online reviews as a signal of the quality of the product (Situmeang, Leenders & Wijnberg, 2014) and aid the online reviews in their purchase decisions (Nielsen, 2012). In addition, Bickart and Schindler (2001) state that consumers have increasingly relied on online reviews for their search of information related to a variety of products.
2.2 PURCHASE INTENTIONS
In this study, we investigate the purchase intentions of consumers, because research on the actual sale of a product or service can only be done afterwards. In addition, Armitage and Christian (2003) argue that attitudes only serve to direct behavior to the extent that they influence the intentions. Armitage and Christian (2003) describes intentions as “a summary of the motivation required to perform a particular behavior, reflecting an individual’s decision to follow a course of action, as well as an index of how hard people are willing to try and perform the behavior”. Schiffmann & Kanuk (2000) argue that the higher the purchase intentions are, the consumers are more likely to buy the product.
2.2.1 RELATIONSHIP BETWEEN ONLINE REVIEWS AND PURCHASE INTENTIONS
As stated in the review about online reviews, consumers have increasingly relied on online reviews for their search of information related to a variety of products (Bickart and Schindler, 2001). Several studies investigate how various characteristics of online reviews influence the buying intentions of consumers. First Park et al. (2007) assume that the number of reviews and consumers who actually have bought the product are related. Zhang, Zhao, Cheung and Lee (2014) argue that consumers have a higher purchase intention with a higher amount of online reviews. Moreover, the (perceived) quantity of reviews positively affects consumers’ willingness to buy (Zhang et al, 2014; Lee, 2009; Park et al, 2007). Also the valence2 of online reviews influence the purchase intention of consumers. A negative review have more impact than a positive review on the buying intention of consumers (Bae & Lee, 2011; Zou et al., 2011). However, East et al. (2008) state that positive online reviews had more impact on brand purchase probability. Studies find a positive relationship (Lin, Fang and Tu, 2010) and a negative relationship (Chevalier and Mayzlin, 2006) between the valence of online reviews and consumers’ purchase intentions. The credibility of the source has an impact on the persuasion and the attitude change (Zie et al., 2011), but no significant result are found that source credibility have an impact on the purchase intentions of consumers (Cheung et al. 2008). In addition,
Bambauer-Sachse and Mongold (2013) find that consumers who have such knowledge are less influenced in their product evaluations by reviews, especially by the negative ones. Dependent on the source credibility, Bambauer-Sachse and Mongold (2013) also find that the effects of negative reviews are even weaker when consumers acquire their knowledge by a highly credible source compared to a less credible source. In other words, high source credibility intensify the knowledge effect. Several moderators like gender or impulsiveness of consumers influence the purchase intention of consumers (Zhang et al., 2014; Zhang et al., 2007). In summary, online reviews are one of the most important eWOM and have a direct influence on the purchase intentions of consumers. However, all these studies assume that the online reviews are honest reviews submitted by real consumers and/or experts.
2.3 MANIPULATED ONLINE RE VIEWS
Liu & Park (2015) state that online reviews allows consumers to search for not only detailed but also reliable information. However, Situmeang et al. (2014) argue that reliable information is hard to find, because the product quality is difficult to evaluate before consumption. This phenomenon creates the possibility for organizations to manipulate the reviews as marketing approach to increase their sales or to hurt the competitors (Dellarocas, 2006; Hu et al, 2011). In addition, reviews can be posted positive about their own products or negative about the competitor products (Dellarocas, 2006). Following previous literature, reviews manipulation can be defined as “vendors, publishers, writers or any third-party consistently monitoring, the online reviews and posting non-authentic
online reviews on behalf of customers when needed, with the goal of boasting the sales of their
products” (Hu, Liu and Sambamurthy, 2011). In figure 1, there is a description of a special published
Figure 1: Article about manipulated reviews on Hotel review sites (www.gov.uk)
“On the 19th of June 2015, BBC came with a special that businesses are ambushing rivals with fake reviews (bbc.com). The Competition and Markets Authority (CMA) started an investigation into businesses that may be paying for endorsements in blogs and other online eWOM forms where the payment may not been made clear to the readers (www.gov.uk). The CMA argue that in addition to the two previously-mentioned ways to manipulate reviews, two other methods are used to mislead consumers: 1. Review sites cherry-picking positive reviews; and 2. Sites allowing businesses to remedy negative reviews, that go unpublished, meaning a complete picture is not clear to review site users”
In addition to this article, a few studies have be done about the manipulation of the reviews. Mayzlin, Dover & Chevalier (2013) argue that they found both manipulation strategies on two major sites for hotel reviews. Moreover, review skeptic offers a tool that uses machine learning to identify fake hotel reviews based on research at Cornell University. Based on this research, Hu et al. (2012) find that 10.3% of the products are subject to online reviews manipulation by examining the writing style. They also argue that consumers are only able to detect the manipulation taking place through ratings, but not through sentiments. Overall, it is still difficult to find out whether reviews for products or services are manipulated. Review manipulation is a new phenomenon and that is why there is limited understanding about how to detect review manipulation (Hu, Bose, Gao and Liu, 2011). In addition, the tendency to believe online reviews vary by online consumers (Sher and Peter, 2009). However, Bambauer-Sachse and Mangold (2013) argue that consumers are more conscious about these marketing approaches of manipulated online reviews (Bambauer-Sachse & Mangold, 2013). That is why this research will investigate the impact of the knowledge about manipulated online reviews on the buying intentions of consumers.
2.3.1 RELATIONSHIP BETWEEN MANIPULATED ONLINE REVIEWS AND PURCHASE INTENTIONS
To explore the relationship between the knowledge about manipulation of the online reviews and the purchase intentions, consumer relationship with the brand or service provider is an important driver to acquire new customers, brand equity, customer retention and overall more profit (Reichheld, 1996; Blackston, 2000; Winer, 2001; Dowling, 2002). Moreover, consumers’ trust and commitment are central components of relationships between the brand and/or provider, and the consumers (Berscheid, 1994). Consumers’ trust encompasses two dimensions: First, the consumers’ belief that the product has the skills, ability and expertise (Singh & Sirdeshukh, 2000). Secondly, the consumers’ belief that the provider is concerned about the welfare and best interests of the consumers (Ganesan, 1994). Especially, the last dimension is a concern for the brand or provider when consumers know that the online reviews are manipulated. With clear and strong information about the manipulation of the reviews, it may be difficult to maintain consumers’ trust. According to the literature, Harris and Goode (2004) find links between consumers’ trust and their purchase intentions. In addition, Nikbin, Marimuthu and Hyun (2013) found that there is a negative relationship between trust and switching intentions. Therefore, according to the literature the following hypothesis may be inferred:
H1. There is a negative relationship between the knowledge about manipulates online reviews and the purchase intentions.
However, the second component of relationships is commitment with the brand. Commitment can be defined as “an enduring desire to maintain a valued relationship” (Moorman et al., 1992). Previous studies has shown significant differences in the relationships between trust and commitment, and the sales of products or services (Garbarino & Johnson, 1999). Commitment has a two component nature: affective and cognitive. Matilla (2004) argue that emotionally bonded customers might feel “betrayed” when a service failure occurs. However, Varela-Neira et al. (2008) find that these emotions only have an indirect effect on overall satisfaction through cognitive evaluations. Therefore, it is very interesting to understand the cognitive commitment through the measure of brand knowledge.
2.4 BRAND KNOWLEDGE
In order to know the definition of brand knowledge, it is necessary to explore the brand name and to define the concept of consumer-based brand equity first. A brand can be defined as “a name, term sign, symbol, or design, or combination of them which is intended to identify the goods and
services of one seller or group of sellers and to differentiate” (Keller, 2001). Keller (1993) defines
consumer-based brand equity as “the differential effect on brand knowledge on consumer response to the marketing of the brand”. According to Aaker (1996), consumer-based brand equity is made of the set of assets and liabilities.
Peter and Olson (2001) state that consumer brand knowledge relates to the cognitive representation of the brand. Further, brand knowledge can be defined in terms of the personal meaning about a brand stored in consumer memory, that is, all descriptive and evaluative brand-related information (Keller, 2003). Moreover, Keller (1993) argue that “Brand knowledge is conceptualized according to an associative network memory model in terms of two components;
brand awareness and brand image”. According to Aaker (1996), the assets and liabilities increases or decreases the value of the product of services when they are linked to the brand. When familiarity with the brand by the consumer rises, they hold favorable, strong and unique brand associations in mind (Keller, 2001). In response to marketing activities of a brand, it is important to understand brand knowledge, because they influence what comes to mind when a consumer thinks about a brand (Keller, 2001). Brand knowledge affects consumer response to current marketing activity (Keller, 2003). Brand awareness and brand image will be explored in the next two paragraphs.
2.3.1 BRAND AWARENESS
Brand awareness can be defined as the level of brand recall and recognition that consumers have of a particular brand and its specific product category (Keller, 2001). Aaker (1996) defines brand awareness as the consumers’ association with a particular product to a certain brand. Keller (2001) uses two constructs to measure brand awareness; Recall and Recognition. Keller (2001) describes brand recall to measure brand awareness as the correct identification of brand given product category
or some other type of probe as cue. The purpose of the measure is to capture “top-of-mind” accessibility of brand in memory. Keller (2001) describes brand recall to measure brand awareness as the correct discrimination of brand as having been previously seen or heard.
Brand awareness can be used as a tactic to let consumers make their decisions faster with high awareness brands than subjects in the non-awareness conditions (MacDonald, Sharp, 2000). Brand awareness seems to be a powerful influence on brand choice in a repeat purchase consumer product context (MacDonald, Sharp, 2000). Previous studies show that brand awareness has a high impact on the outcome variables (Huang & Sarigöllü, 2012; Homburg, Klarmann, Schmitt, 2010). Other studies show that the higher is the brand awareness, brand loyalty will increase and the higher is the purchase intention (Chi, Yeh & Yang, 2009). Moreover, Malik and Ghafoor (2013) argue that brand awareness has a strong positive association with purchase intention. The investigation of the role of brand names in customization decisions by Pingjun Jiang (2004) suggests that brand names can have more impact in choice-making by determining the extent of the perceived preference match. Especially, when relevant information is missing, consumers will rely more on the prior information about the brand (Ratchford 1982). However, all these relationship between brand awareness and another outcome variable are based on general information. In this study, we research the consequences between a negative variable (manipulation) and the purchase intention, and the moderating role of brand awareness on this negative relationship.
2.3.2 MODERATING ROLE OF BRAND AWARENESS
From marketing science, no literature was found to explore the effects of the knowledge about manipulated online reviews and brand awareness. However, from a psychological perspective on relationships, Mattila (2004) argue that consumers may feel betrayed by the brand or service provider, when they know that the online reviews are manipulated. The effects of this betrayal can be explored by the concept of brand commitment (Dick and Basu, 1994). The concept of brand commitment have two key constructs: affective and cognitive commitment (Fournier, 1998; Ahluwalia et al., 2001). Previous studies show that cognitive commitment results in defense motivation with negative
information (Eagly and Chaiken, 1995) and exhibited lower levels of attitude change than low commitment (Ahluwalia et al., 2000). However, we propose that the manipulation of the online reviews as a marketing approach is a phenomenon that can’t be counter-argued or discounted. According to Bitner et al. (1990), “the betrayal effects” might magnify the negative consequences on loyalty. From the affective commitment perspective, Mattila (2004) argue that emotionally-bonded consumers reacts in sharp decrease in their attitudes and loyalty to the brand. Conversely, consumers with a low level of emotional bonding with the brand are more “forgiving”. In this specific case, we suspect that the knowledge about the manipulation of the online reviews has a higher level of negative impact on the purchase intentions of consumer for consumers with a higher level of brand awareness. Conversely, we suspect that for consumers with a lower level of brand awareness, the knowledge of manipulation of the online reviews has a weaker negative impact on the purchase intentions. In addition to the literature, the following hypothesis may be inferred:
H2. The negative relationship between the knowledge about manipulated online reviews and the purchase intentions of consumers is moderated by brand awareness, so that this relationship is
stronger for higher values of brand awareness than for lower values of brand awareness.
2.3.3 BRAND IMAGE
Brand image can be defined as the set of multiple mental associations, such as different types, favourability, strength and uniqueness, with the brand which influence the buyers by their purchase decision (Keller, 2001). Aaker and Biel (1993) show that brands can evoke feeling and associations. Keller (1993) argues that it is necessary that a brand node has been established in the consumer’s memory to create brand image. Aaker and Biel (1993) state that brand image has three components which are related to the personality and character of the brand: the image of the provider, the user and the product itself. When consumers face the brand, the symbol of the brand are automatically accessed from the consumers’ memory (Aaker and Biel, 1993). Ike-Elechi & Zhenzhen (2009) argue that brand image can positively influence customers’ loyalty to a market offering and possibly boost customer commitment. Ike-Elechi & Zhenzhen (2009) also suggest that good brand image influence customer
perceived quality, enables satisfaction and should also influence to a greater degree the extent to which customers are willing to express commitment to such offering for a sustainable profit. However, these previous studies are based on the fact that the marketing is done on the basis of general information. This study focus on the fact that consumers already know that the online reviews are manipulated. Therefore, we investigate the role of brand image on the negative relationship between the knowledge about manipulation of the online reviews and the purchase intentions.
2.3.4 MODERATING ROLE OF BRAND IMAGE
Jalilvand and Samiei (2012) argue that online reviews is one of the most effective factors to influence brand image and purchase intentions. Depending on brand image, manipulation of the online reviews can lead to negative consequences on brand image. For example, Wu et al. (2011) investigate the effects of service quality on brand image and show that service quality has a direct and positive effect on brand image. In this study, the results show that a lower level of service quality leads to a lower level of brand image. In addition, Nam-Hyum (2013) find that negative brand information has negative impacts on the brand evaluation and the purchase intentions. Moreover, Bambauer-Sachse and Mangold (2011) find support for the detrimental effect of negative online reviews on consumer-based brand equity. According to previous studies, we propose that the manipulation of the online reviews as a marketing approach is a phenomenon that can’t be counter-argued or discounted, because consumers may feel betrayed (Mattila, 2004). This “betrayal effect” might magnify the negative consequences of brand image (Bitner et al., 1990). According to the brand relationship, described in the paragraph about brand awareness, we suspect that consumers with emotional bonding reacts in sharp decrease in their attitudes and loyalty to the brand. Compared to the moderator variable brand awareness, the decrease will be sharper, because of direct impact of the negative information on the brand image (Nam-Hyuam, 2013). Conversely, consumers with a low level of emotional bonding with the brand are more “forgiving”.
Therefore, we propose that the knowledge about the manipulation of the online reviews has a higher level of negative impact on the purchase intentions of consumers for consumers with a higher
level of brand image. Conversely, we propose that for consumers with a lower level of brand image, the knowledge of manipulation of the online reviews has a weaker negative impact on the purchase intentions. In addition to the literature, the following hypothesis may be inferred:
H3. The negative relationship between the knowledge about manipulated online reviews and the purchase intentions is moderated by brand image, that this relationship is stronger for higher values
3. CONCEPTUAL MODEL
Figure 2 provides an overview of the conceptual framework of this study. In the following chapter, an overview of the hypothesis are mentioned.
Figure 2: Conceptual Framework
3.1 HYPOTHESIS
The following hypotheses will be tested:
H1. There is a negative relationship between the knowledge about manipulates online reviews and the purchase intentions.
H2. The negative relationship between the knowledge about manipulated online reviews and the purchase intentions of consumers is moderated by brand awareness, so that this relationship is
stronger for higher values of brand awareness than for lower values of brand awareness.
H3. The negative relationship between the knowledge about manipulated online reviews and the purchase intentions is moderated by brand image, that this relationship is stronger for higher values
4. RESEARCH DESIGN
To investigate the impact of brand knowledge on the relationship between the knowledge about manipulated online reviews and purchase intentions, data will be collected. This chapter describes how data was collected. In the first part of this chapter, the sample of the study is described. In the second part, the research design is described including explanation about the questionnaire that will be used and the measures. In the last part of this chapter, the procedure about how data will be collected is described.
4.1 SAMPLE
The research reflects the philosophy of positivism from a deductive approach. This quantitative research will be carried out using a cross-sectional experimental survey design, which is a study of a particular phenomenon at a particular time, because of the time horizon of the study. Thereby, it is useful to test hypotheses and generalize the findings. The online survey will be set up with the program Thesistools and the questionnaire will be distributed digitally to supervisors and their employees of different companies. The approached population for this study will be mainly employees of different workplaces in Dutch workforce. The main focus group for the population will be the company Boval Group BV. Through contacting individual workers in the company and different supervisors, the respondents can be selected. The population within the company is quite large and diverse, so the aim is for purposive heterogeneous sampling. However, because of the difficulty to getting access to the different departments and finally getting response from the employees, convenience sampling is also used. Thereby, more respondents can be selected through contacting other supervisors from different companies. Because of the sampling technique, the sample will likely be diverse. However, this might not be a problem to this study, because this research strives for as many respondents as possible. Considering that there is limit, we strived to more than 200 respondents, because according to the rule of thumb within pragmatic limit of respondents’ time availability this would be a good sample size.
At the end of the questionnaire respondents are asked for their demographics, such as gender (nominal variable), age (ratio variable) and level of education (nominal variable). For all the other questions and/or statements validated Likert scales will be used on a 7-point scale (completely disagree – completely agree) at interval level. All items used in the questionnaire were derived from English studies. However, the questionnaire would be translated into Dutch, because the target group are employees within Dutch companies. So this will lead to a better understanding in the respondents. To assure that the content of the items remains unchanged, the translated Dutch items were back translated into English by a third person. In the Appendix 2, the Dutch and English survey is included.
4.2 RESEARCH DESIGN
To be able to identify and explain the effect of manipulation of the online reviews on the purchase intention, and to test whether the effects differ from the level of brand awareness and brand image, a quantitative research through an online experimental survey was most appropriate. The online survey offers easy access to a diverse population, cost savings and the ability for participants to access in their familiar situation increasing the external validity (Reips, 2000).
A combination of an online survey and a 2 x 2 x 2 factorial experimental design (Figure 3), was used to be able to test the hypotheses that have been developed. The experimental design aims at predicting the outcome by introducing a change of the preconditions (the knowledge about the manipulation of the online reviews), which is reflected in the predictor variable (online reviews). The change in the predictor is generally hypothesized to result in a change in the outcome variable (purchase intentions). As shown before, the independent variables in this study is (the knowledge about the manipulation of) the online reviews. Before every separate survey, we determine the knowledge about the manipulation of the online reviews by omitting or introducing that the online reviews are manipulated. For the data analysis, we insert a column in the data file with the knowledge about the manipulation. Two values are used; 0 is the value of no knowledge and 1 is the value of the knowledge about the manipulation. The dependent variable is the purchase intention of the consumers. To determine the relationship between the knowledge about manipulation of the online
reviews and the purchase intention, three control variables are used: Gender, age and level of education. The moderator variables are brand awareness and brand image. Every survey includes four items per variable to measure the value of both individual variables. After collecting the data, we determine the means of both variables and recode the means of both variables into mean centered different variables. After recoding this new variables, we determine whether this new variables are negative or positive and recode this mean centered variables into new different variables. This new variables have two values; 0 is the value for low level (negative compared with the mean) of brand awareness and low level of brand image, and 1 is the value for high level (positive compared with the mean) of brand awareness and high level of brand image. Figure 3 includes an overview of the factorial design of this study.
Figure 3: 2 x 2 x 2 Factorial Design
High level of brand awareness Low level of brand awareness
High level of brand image Low level of brand image High level of brand image Low level of brand image No knowledge about manipulation
Treatment 1 Treatment 3 Treatment 5 Treatment 7
Knowledge about manipulation
Treatment 2 Treatment 5 Treatment 6 Treatment 8
4.2.1 SURVEY PROCEDURE
Before distributing the questionnaire, a pretest is needed to determine the level of brand awareness. Randomly selected brands
(Reebok, Nike, Adidas, Asics, Dutchy, Lotto, Osaga and Le Coq Sportif) are tested and then two brands are selected with the highest value and the lowest value of brand awareness for using in the questionnaire. A
validated Likert scales will be used on a 7-point scale (low – high) at interval level. On October 16, 2015 the pretest was distributed on the department Income insurance of the Boval Group. A total of 15 respondents have made the pretest. In the Appendix 1 the data of the pretest is included. Table 1 represents the mean scores of the brands form the pretest. Taking into account the survey design, we look to the highest and lowest scores, and then we search for the brands who offer the same model of a product. For the online experiment, two different brands with the same product were chosen. The two brands are Nike and Osaga, and the product is a sport sneaker, namely Nike Roche Run trainers and Osaga man sport sneaker. Both shoes have the same product descriptions and are the same models.
The experimental survey consisted of 4 parts (Appendix 2). The first part of the experimental survey contained some general questions about online reviews to determine whether the online reviews are related with the purchase intentions of the consumers. By the second part, we showed the product of Nike’s Roche run trainer and the online reviews. The second part contained questions to determine the brand awareness, the brand image and the consumer’s purchase intentions about the product of Nike. At the beginning of the third part, we showed the product of Osaga’s sport sneaker with the same online reviews compared to the product of Nike. Further the third part contained questions to determine the brand awareness, brand image and the consumer’s purchase intentions about that product. The last part of the experimental survey is the part whereby we introduced an article about the manipulation of the online reviews and we indicate that both Nike as Osaga were guilty of deliberately manipulating the online reviews for both articles. Thereby, the last part consisted of several questions about trust, commitment and loyalty to determine the value of the knowledge about the manipulation of the online reviews, and questions to determine the purchase intentions of the consumers. We finished the experimental survey with three questions about the respondent’s age, gender and level of education.
4.2.2 MEASURES
Firstly, every survey starts with general questions about the online reviews. Therefore, we research whether the online reviews are used by consumers before having the intentions to purchase the product. The use of the online reviews will be determined by 4 items with questions about checking the online reviews, the importance of the reviews and the valence of the reviews. In the data analysis, we start with the research about the use of the online reviews by consumers.
Secondly, to determine whether there is a negative relationship between the knowledge about manipulation of the online reviews and the consumer’s purchase intentions, we research the purchase intentions of both preconditions, namely no knowledge about the manipulation and knowledge about the manipulation of the online reviews. Therefore, we use two variables; the independent variable “Manipulation” and the dependent variable ´purchase intentions´. The independent variable “Manipulation” is the insert new variable of the preconditions, namely 0 as the value for no knowledge about the manipulation and 1 as the value for the knowledge about the manipulation. The dependent variable “purchase intentions´ is the new recoded variable from the different items used in the survey to measure this variable after doing a reliability analysis. In this study, we recode the data into new different variables from the two different preconditions, namely the study without knowledge about the manipulation of the online reviews (TotGenPI) and with the knowledge about the manipulation of the online reviews (TotManPI). After recoding this variables, we recode this variables into the new dependent variable ‘purchase intentions’ and call this total variable of the purchase intentions “TotPI”. With this new recoded variables, we research whether there is a negative relationship between the knowledge about the manipulation of the online reviews and the purchase intentions. To study the negative relationship, the following three control variables are used: Gender, Age and Level of education.
As last, we research the moderating effect of brand awareness and brand image in the negative relationship between the knowledge about manipulation of the online reviews and the purchase intentions. The moderating variable ‘brand awareness’ will be derived from the four
questions about brand awareness in part two and three per product. After reliability analysis, we recode the four items per product into a new different variable ‘brand awareness’ and call this moderator variable ‘TotBA’. For the other moderator variable ‘brand image’, we use the same method and call this variable ‘TotBI’.
4.3 PROCEDURE DATA COLLECTION
Data was collected via an online experimental survey. The reasons are explained in Section 4.2 Research Design. It has been created and hosted on www.thesistools.com in Dutch. We chose Thesistools for our experimental survey because of the user friendliness of the program and the previous experience of the researchers. Thesistools is very easy to work with and is very flexible in the type of questions. However, one disadvantage of using Thesistools is that the data obtained from Thesistools had to be converted to the file used by SPSS.
4.3.1 PILOT STUDIES
Prior to the main study, two pilot studies were conducted. With the pilot study we were able to evaluate the study and assure a higher quality of the experimental survey. The first study was executed offline with two friends (both University Master degree) and it was done on October 13, 2015. The reason for the offline pilot study was to see whether the questions were complicated and the structure of the survey was unclear. From the feedback, some questions were deleted. On October 20, 2015 the online pilot study was executed with two other different participants (One HBO degree and one PhD student). The online pilot study served to see whether there were any critiques and/or improvements with respect to the experimental study. Both participants pointed out that the questions among the online reviews have to be seen on one page so that participants do not have to scroll down. Besides this, there were no other critiques regarding the survey.
4.3.2 MAIN STUDY
As mentioned above, the data has been collected through an online experimental survey which was distributed on November 5, 2015 and it was deactivated on December 23, 2015. The data
was created and hosted on www.thesistools.com. Besides the online experimental survey, we distributed several paper-and-pencil formats and 124 respondents filled out this format. The collected paper-and-pencil formats we have manually filled out in the online experimental survey.
The main focus group for the population will be the company Boval Group BV. Through contacting individual workers in the company and different supervisors, the respondents can be selected. The population within the company is quite large and diverse, so the aim is for purposive heterogeneous sampling. However, because of the difficulty to getting access to the different departments and finally getting response from the employees, convenience sampling is also used. Thereby, more respondents can be selected through contacting other supervisors from different companies. Because of the sampling technique, the sample will likely be diverse. Nevertheless, in total 660 participants had started the survey of which in the end 592 were usable for analysis.
On November 5, 2015 the survey was distributed and it was deactivated on December 23, 2015. From the 660 respondents that started filling out the questionnaire, 592 respondents fully completed the questionnaire. From the 592 respondents, 124 respondents filled out the paper-and-pencil format. Taking all 592 respondents together (Mage = 38.92, SDage = 11.305, age-range: 19-81 years) 324 were male (54.7%) and 268 were female (45.3%). Most of the participants (80.5%) were the ages between 21 and 50. The sample covered a broad range of educational backgrounds. 8.1% of the respondents had only completed High School and 27.7% of the respondents have Lower vocational education (MBO). Most of the participants have an education level of HBO (42.6%). At last, 20.9% of the respondents have University master degree background and 0.7% filled in “Others”.
5. RESULTS AND ANALYSIS
In this chapter preliminary analysis was done before testing the reliability of the scale variables and the testing the hypotheses developed by doing further analysis in SPSS. The preliminary analysis contained dealing with missing values, recoding, descriptive, kurtosis, skewness and normality check.
5.1 PRELIMINARY STUDY
In total 660 questionnaires has been filled in. However, not all these questionnaires are complete, because in the survey the respondents were able to stop the survey. There we did a check of frequencies to examine if there were any errors in the data. The amount of missing data was < 10% for all variables. To deal with this missing data, we excluded cases listwise. This means that only the cases without any missing data were analysed.
Items that are phrased so that an agreement with the item represents a low level of construct were recoded into different variables to assure there were no counter-indicative items. This means that rManPI3 and rManPI5 have been recoded and now represents ManPI3 and ManPI5. In the data file
we have to use dummy variables for two variables, namely gender and education. With dummy coding we find a way of representing categorical variables with multiple categories through several dummy variables. For gender, we use 1 dummy with 0 for female and 1 for male. For the second variable education, we use three dummy variables, because this variable has four categories. In this study, we use University degree as the reference group and create three dummies, namely Dummy1_HighSchool, Dummy2_MBO and Dummy3_HBO. As last, we insert the new variables
manipulation, level of brand awareness, brand image and purchase intentions. If the survey was
based on general online reviews (without the knowledge of manipulated online reviews), we insert 0 and for the surveys based on the knowledge about the manipulation we insert 1 in a new column called ‘Manipulation’.
By the Kolmogorov-Smirnov test, all variables are tested on normality. All the variables have a significant normal distribution (p<.05). Most of the other individual items of the variables have no or a moderate skewness and/or kurtosis. The items GenBA2 scores a substantial skewness and a moderate kurtosis. The following items score a moderate skewness and a substantial kurtosis: AlgOR3, GenBA1, GenBI1, GenBA4, GenPI6, GenPI7, ManOR1 and ManOR3. But “with reasonable large, skewness will not make a substantive difference in the analysis” (Tabachnick & Fidell, 2001). “Kurtosis can result in an underestimate of the variance, but the risk is also reduced with a large sample” (Tabachnick & Fidell, 2001).
5.2 RELIABILITY, SCALE MEANS AND CORRELATIONS
For this study, averages of variables with a multi-item scale has to be computed. Therefore, a reliability analysis of the scales had to be checked. The Cronbach’s Alpha, which represents the estimator of the internal consistency, has been tested.
The first deleted item is ManPI1, because the item has a low correlation with the total score of the scale (.222) and if the item is deleted would affect the reliability score. After doing the reliability analysis without the deleted item, the Cronbach’s Alpha score
is 0.699 and all items have a good correlation with the total score of the scale and no items would substantially affect if one of the items is deleted. As exhibit in table 2, most variables have a Cronbach’s Alpha >.7 which indicates a high level of internal consistency. Only TotManPI have a Cronbach’s Alpha lower .7 (0.699). All of the corrected item-total correlation indicate that all items have a good correlation with the total score of that particular scale (all above .30). Also, none of the individual items would substantially affect reliability of one is deleted of the particular scales.
As the final step, new variables as a function of existing variables were created for hypothesis testing. The mean of all items that was used to describe a variable are calculated and we set up the new variables TotOR and TotPI. Both variables are all the score of both the questions of the knowledge about the manipulated and the general online reviews. In table 3, the means and standard deviations of all variables in the research with online reviews are exhibited. To research the relationship between the four variables per condition, a correlation matrix is compiled from SPSS.
From the correlation carried on the general online reviews, two of the three variables have a significant correlation with the purchase intention (TotPI). Brand awareness (TotBA) and brand image (TotBI) have a correlation with each other at a significance of 0.01 level, whereby brand image (.290)
has the highest correlation and brand awareness (.287) a few lower correlation with the purchase intentions of consumers. The general variable online reviews (TotOR) has no significant correlation with the purchase intentions of consumers. Further, there is a significant correlation between the online reviews and brand image at the 0.05 level (0.081) and between brand image and brand awareness at the 0.01 level (0.810).
5.3 ANALYSIS RESULTS
First, we investigate the general use of the online reviews. With the descriptive statistics we calculate the means. All items have a significant mean above 5 on a Likert scale from 1 till 7. The question about checking the online reviews when buying a product scores 5.0405. The question about the importance of using the online reviews scores 5.0743. The questions about using positive reviews
(5.2365) and using negative reviews (5.0811) when buying a product online also score above 5. After doing the reliability analysis (Cronbach’s Alpha=.874), we recode the items into a new variable “TotAlgOR” and research whether the general information about the online reviews effects the purchase intentions of consumers. We found that the online reviews have no significant effect on the purchase intentions of consumer (p>.05).
After analysing the general use of the online reviews, we investigate whether there is a negative relation between the knowledge about manipulation of the online reviews and the purchase intentions with a linear regression analysis. Hierarchical multiple regressions were performed to investigate the ability of the knowledge about manipulated online reviews to predict levels of purchase intention, after controlling for gender, age and level of education. Further, we investigate whether the effect of the knowledge about the manipulated online reviews on the purchase intentions of consumers depends on the two moderator variables brand awareness and brand image. Therefore we use a regression analysis with Process.
Hypothesis testing – hierarchical regression (General online reviews)
To examine the linear relation between the online reviews (TotOR) and the purchase intentions of consumers (TotPI), we use a regression analysis. With this analysis we understand the effects of the online reviews on the purchase intentions based on the theoretical framework and we identify a linear combination of the online reviews to predict in the best way the value of the purchase intentions.
In the first step of the hierarchical multiple regression, three predictors were entered: gender, age and level of education. This model is statistically not significant F (6, 585) = 1.154 (p >.05) and explained 1.2% of the variance in purchase intention. After entry of the online reviews (TotOR) the model is statistically significant and the total variance explained by the model as a whole was 6.2%, F (7, 584) = 31.246; (p <.001). The introduction of the online reviews explained additional 5.1% variance in purchase intentions, after controlling for gender, age and education (R2 change = .051; F
(1, 584) = 5.504; p <.001). In the final model, two out of four predictor variables were statistically significant, namely age (β= -.092, p<.05) and the manipulation of the online reviews (β= .224, p<.001). Table 4 exhibits the hierarchical regression model of purchase intentions.
Hypothesis testing – testing the moderating effect of brand awareness (knowledge about manipulated online reviews)
Further, the moderating effect of brand awareness was analysed. Table 5 shows relevant results of the analysis. By adding brand awareness as a moderator the model is able to explain 16.83% of the variance in the purchase intention of consumers. This is an additional 10.63% compared to the previous model. Further it could be shown that for a level of confidence of 99.9% there is a significant (P=.0000; p>.001)
moderating effect of brand awareness (β=-.2569, p>.001) on the relation between the knowledge about manipulated online reviews and the purchase intentions of consumers. The moderating effect has a significance for two levels of brand awareness (p<.05), namely low and moderate. For the high level of brand awareness (p>.05), there is no significance. The interaction leads to an increase of R2 with 3.58%. The interaction leads to an increase of R2 with 3.58%. Table 6 shows the interaction variables. As table 5 and table 6 show, there is a significant moderation effect between the level of brand awareness and the knowledge about the manipulation of the online reviews on the purchase intentions of consumers. We found that the purchase intentions of consumers with a low level of brand awareness are rising faster than the consumers’ purchase intentions with a high level of brand awareness. Figure 4 illustrates this effect.
Hypothesis testing – testing the moderating effect of brand image (knowledge about manipulated online reviews)
Finally, the moderating effect of brand image was analysed. Table 7 shows relevant results of the analysis. By adding brand image as a moderator the model is able to explain 17.65% of the variance in the purchase intention of
consumers. This is an additional 11.45% compared to the previous model. Further it could be shown that for a level of confidence of 99.9% there is a significant (P=.0000; p>.001) moderating effect of brand image (β =-.3276, p>.001) on the relation between the knowledge about manipulated online
reviews and the purchase intentions of consumers. The moderating effect has a significance for two levels of brand image (p<.05), namely low and moderate. For the high level of brand awareness (p>.05), there is no significance. The interaction leads to an increase of R2 with 4.23%. Table 8 shows the interaction variables. As table 7 and table 8 show, there is a significant moderation effect between the level of brand image and the knowledge about the manipulation of the online reviews on the purchase intentions of consumers. We found that the purchase intentions of consumers with a low level of brand image are rising faster than the consumers’ purchase intentions with a high level of brand image. Figure 5 illustrates this effect.
6. CONCLUSION AND DISCUSSION
As we already know from previous literature, consumers have increasingly relied on online reviews for their search of information related to a variety of products (Bickart and Schindler, 2001). From this study, we found no statistically significance that the use of the online reviews effects the purchase intentions of consumers. However, for using the online reviews as search of information, the general information about the online reviews scores significant mean values above 5 for all four items used in the survey. Therefore, we confirm the statement of Bickart and Schindler (2001) that consumer relied on online reviews for their search of information related to a variety of products. However, this phenomenon creates the possibility for organizations to manipulate the reviews as marketing approach to increase their sales or to hurt the competitors (Dellarocas, 2006; Hu et al, 2011). In addition, the tendency to believe online reviews vary by online consumers (Sher and Peter, 2009). Bambauer-Sachse and Mangold (2013) argue that consumers are more conscious about these marketing approaches of manipulated online reviews (Bambauer-Sachse & Mangold, 2013). This study investigates the effect of the knowledge about manipulation of the online reviews on the purchase intentions and also investigates the role of brand knowledge in this relationship. As illustrated in Table 9, hypothesis 1 is rejected and hypotheses 2 and 3 are partial supported.
As Hu, Bose, Gao and Liu (2011) already stated, there is a limited understanding about the phenomenon of review manipulation. From a psychological perspective we assume that, when the consumers know that the online reviews are manipulated, this knowledge has a negative effect on the purchase intentions. However, from the results of our study, the knowledge has actually a significant positive effect on the purchase intentions of consumers. Therefore, we have to reject hypothesis 1. The results of our study are not consistent with the literature about trust. According to Berscheid (1994), consumers’ trust encompasses the belief that the product has the, ability and expertise (Singh & Sirdeshukh, 2000) and the belief that the provider is concerned about the welfare and best interests of the consumers (Ganesan, 1994). In our literature review, we suspect that this last dimension is a concern to trust the online review and that this concern has a negative effect on the total trust of buying the product online. However, related to the results, we found that the knowledge about the manipulation actually has a positive effect on the intentions to buy a product. A possible explanation is that the consumers’ trust is not so much relied on the online reviews, but more on the product specifications. Therefore, we refer to the statement of Bickart and Schindler (2001) that consumers have increasingly relied on online reviews for their search of information. Another explanation is that
consumers are more conscious about these marketing approaches (Bambauer-Sachse and Mangold, 2013). In addition, Shiv, Edell and Payne suggest that under certain condition, negative framing may generate tactics-related cognitions, thereby diminishing the impact of negative information. Therefore, the effect of the knowledge of the manipulation on the purchase intentions is not negative, because of the fact that consumer are more and more conscious about the manipulation of the online reviews. However, the positive effect on the purchase intentions is still hard to explain. The fact that organizations could manipulate the online reviews while consumers know this phenomenon and that this will lead to a higher purchase intentions of the consumers, could be explained by the theory of avoiding information. According to Sweeny et al (2010), consumers avoid information (i.c. the knowledge of the manipulation of the online reviews) if they expect bad news. If this is the case, the negative publicity can increase purchase likelihood and sales by increasing product awareness (Berger, Sorensen, Rasmussen, 2010).
Furthermore, we were interested in showing whether there was a moderating effect of brand awareness on the (negative) relation between the knowledge about manipulated online reviews and the purchase intentions of consumers. Therefore, we expected that this relationship is stronger for higher values of brand awareness than for lower values of brand awareness. Our findings partial support hypothesis 2. First of all, we found a moderating effect on the relation between the knowledge about manipulation of the online reviews and the purchase intentions. However, the results show that for both levels of brand awareness, there has been a positive effect on the purchase intentions of consumers when consumers know that the online reviews are manipulated. In our study, we found that there is a stronger positive relationship between the manipulation and the purchase intentions for a low level of brand awareness compared to a high level of brand awareness. The results are not consistent to the literature. First of all, we propose that the betrayal effect might magnify the negative consequences on loyalty which leads in a sharp decrease in consumers’ attitudes and loyalty to the brand. Thereby, consumers with a low level of emotional bonding with the brand are more forgiving. An explanation for our opposite results could be that for a high level of brand awareness, consumers’ commitment with the brand results in defense motivation with negative information (Eagly and
Chaiken, 1995) and exhibited lower levels of attitude change. On the other hand, low commitment consumers will exhibit a greater amount of attitude change in response to negative information (Ahluwalia, Burnkrant and Unnava, 2000). Mostly, there is a negativity effect, but in this case we would refer to better bad publicity than no publicity.
Finally, we were interested in showing whether there was a moderating effect of brand image on the (negative) relation between the knowledge about manipulated online reviews and the purchase intentions of consumers. Therefore, we expected that this relationship is stronger for higher values of brand image than for lower values of brand image. Our findings partial support hypothesis 2. First of all, we found a moderating effect of brand image on the relation between the knowledge about manipulation of the online reviews and the purchase intentions. Compared to the moderating effect of brand awareness, there has been a positive effect on consumers’ purchase intention when consumers know that the online reviews are manipulated. We also found the same results compared to brand awareness that there is a stronger positive relation between the manipulation and the purchase intentions for a low level of brand image compared to a high level of brand image. These results are also not consistent to the literature. However, for this result we would also refer to the theory of avoiding information if consumers expect bad news (Sweeny et al, 2010) and the greater attitude change for low commitment consumers (Ahluwalia, Burnkrant and Unnava, 2000).
