The Subjectively Objective Consumer: How Perceived Electronic Word-of-Mouth Quality and Brand Commitment Affect Behavioural Intentions

(1)

The Subjectively Objective Consumer:

How Perceived Electronic Word-of-Mouth Quality

and Brand Commitment Affect Behavioural

Intentions

Vincent Pool

S2356821 – V.B.Pool@student.rug.nl

B170807315 - V.Pool2@newcastle.ac.uk

Dissertation

MSc Advanced International Business Management & Marketing

University of Groningen

Newcastle University Business School

Supervisors

Dr. S.R. Gubbi - University of Groningen

Dr. A. Javornik - Newcastle University Business School

(2)

2

ABSTRACT

This study investigates the impact of perceived quality of both positive and negative eWOM on behavioural intentions. Both positive and negative eWOM are incorporated in this study, because the effects of eWOM are inherently dependent on the valence of its content. The perceived quality of eWOM is examined by using online consumer reviews as perceived by other consumers. In addition, the moderation effect of brand commitment on the impact of perceived review quality is examined. In order to expose the relationships between these constructs, this study is placed within a broader theoretical framework and explained through adoption of the Information Adoption Model.

A quantitative research design is used within this study, yielding a sample of 258 respondents. The findings are partly in line with the expectations, proving that a higher perceived review quality has a positive effect on consumer behavioural intentions with regard to positive reviews, and a negative effect with regard to negative reviews. Furthermore, brand commitment is found to alleviate the detrimental effects of negative reviews with a high perceived quality because consumers were found to significantly discount negative messages about their favourite brands. However, contrary to expectations and theory, brand commitment is also found to weaken the positive effect of positive reviews with a high perceived quality. Besides, the moderation effect of brand commitment on positive reviews is found to be stronger than the respective effect on negative reviews. Overall, the findings of this research provide relevant academic and managerial implications, including management eWOM strategies, tactics for the development of brand commitment as a tool to counter negative eWOM effects and future research avenues.

(3)

3

ACKNOWLEDGEMENTS

This Master dissertation marks the end of my study career and the completion of the Master Advanced International Business Management and Marketing at the University of Groningen and the Newcastle University Business School.

First, I would like to thank my supervisors dr. S.R. Gubbi from the University of Groningen and dr. A. Javornik from the Newcastle University Business School for their feedback and advice during the development of this research. Thank you both for your support and this learning opportunity.

Also, I would like to thank all the people who I annoyed with my surveys and who took the time to fill them in. A special acknowledgement is in place for those who snowballed it on. You are part of this research, as you made it possible!

(4)

4

LIST OF TABLES

Table 3.1: Overview of the Experiment Conditions ... 26

Table 3.2: Overview of the Variables Used in This Study ... 31

Table 4.1: Total Socio-Demographic Sample Characteristics ... 35

Table 4.2: Sample Characteristics Per Condition Group (Socio-Demographic) ... 35

Table 4.3: Factor Loadings from the Factor Analysis ... 38

Table 4.4: Factor Loadings from the Factor Analysis of Brand Commitment ... 38

Table 4.5: Factor Loadings from the Factor Analysis of Behavioural Intentions ... 38

Table 4.6: Descriptive Statistics, Cronbach’s Alpha and Correlations ... 39

Table 4.7: Overview of the Manipulation Analysis and Levene Test ... 40

Table 4.8: Overview of Manipulation Tests ... 42

Table 4.9: Hierarchical Multiple Regression Results for Behavioural Intentions ... 44

Table 4.10: Overview of Hypotheses Testing Results ... 47

Table 8.1: Perceived Quality Pre-Test for All Investigated Stimuli ... 69

Table 8.2: List of Utilized Measures for Variables Under Investigation ... 70

Table 8.3: Hofstede Individualism Scores per Respondent Country ... 71

Table 8.4: Descriptive Statistics Pre-Regression ANOVA ... 72

Table 8.5: Pre-Regression 2-Way Analysis of Variance ... 72

LIST OF FIGURES

Figure 2.1: Example of Amazon.com eWOM System ... 14

Figure 2.2: Variables of Interest and Corresponding Hypotheses ... 23

Figure 8.1: Normality Plots for Positive Sub-Sample ... 73

(8)

8

1. INTRODUCTION

The dawn of the internet age has enabled consumers to make more educated choices in their selection of brands faster than ever before. Web 2.0 has enabled consumers to make use of a vast and easily accessible array of tools like consumer review sites, discussion fora and social media to share experiences and to become informed of such experiences by other consumers (Gupta & Harris, 2010; Hennig-Thurau, Gwinner, Walsh, & Gremler, 2004). Word-of-mouth (WOM) is commonly known to be more influential than advertising due to a higher trustworthiness, thus often becoming the dominant factor in the decision making of consumers (Godes & Mayzlin, 2004). The internet has allowed consumers to not only obtain WOM from their immediate surroundings, but to tap into a geographically dispersed pool of people that share their opinions online (Cheung & Lee, 2008). As a result, electronic word-of-mouth (eWOM) is considered to be an issue of high strategic importance in the modern marketing mix (Bickart & Schindler, 2001; Kumar & Benbasat, 2006; Zhang, Craciun & Shin, 2010).

(9)

9

communications may help to explain the inconsistencies in literature. However, existing research has limited itself to identifying the direct influence of eWOM, while neglecting what exactly makes eWOM messages influential in the first place (Filieri, 2015). Hence, this study proposes that the perceived quality of online consumer reviews, as one prominent form of eWOM, influences the behavioural intentions of consumers. Drawing on the information adoption theory (Sussman & Siegal, 2003), the perceived quality of a message is known to directly influence the level of information adoption by a consumer receiving it. Consequently, this is likely to affect the behavioural effects of online reviews. Thereby, the first contribution of this research is to provide empirical evidence for the impact of perceived eWOM quality on consumer behavioural intentions for both positive and negative eWOM messages.

(10)

10

This research sets out to address the identified gaps in literature by exposing the mechanisms underlying eWOM. Therefore, thefollowing research question is proposed:

‘To what extent does the perceived quality of positive and negative eWOM affect behavioural intentions, and to what extent does brand commitment impact this

relationship?’

In sum, the primary objective of this research is to provide insight into how consumers react to both positive and negative online reviews about brands they are committed to, dependent on the perceived quality of these reviews. The survey experiment employed in this study focuses on one particular type of eWOM: online, consumer-generated reviews. This type of eWOM is especially suitable because of the recently increasing relevance, popularity and impact of online review platforms (Deloitte, 2018).

(11)

11

2. THEORETICAL BACKGROUND

2.1 Electronic Word-of-Mouth (eWOM)

The existence of online platforms has provided virtual avenues for consumers to share their opinions about products with each other via the internet. These avenues have resulted in a new wave of word-of-mouth (WOM) communication, now widely known as eWOM. In the field of marketing, WOM communication has always received considerable attention from scholars, who recognize its importance as a tool that is more effective in changing consumers’ attitudes and behaviours than traditional advertising (Bambauer-Sachse & Mangold, 2011). Traditionally however, WOM was restricted to conversations “over the backyard fence” with friends and/or relatives (King, Racherla & Bush, 2014). This changed with the dawn of Internet 2.0, when eWOM was born and its participants engaged in brand-related communication with a network of complete strangers in online communities (Gottschalk & Mafael, 2017). A much larger amount of contributors and audiences is present in online communities than is the case in traditional WOM. Furthermore, conversations exist in traceable, written form on platforms that are not bound by time and location (Cheung, 2009). As a result, product-related and opinionated conversations suddenly became much more visible for consumers, marketers and scholars alike. Online opinions are of particular interest for retailers and website providers because online technologies allow for close monitoring of information for marketing purposes (Burke 2002; (Purnawirawan, de Pelsmacker & Dens, 2012) .

(12)

12

reviews contain favourable information regarding the brand and encourage its purchase, while negative reviews contain unfavourable information and have the opposite effect (East et al., 2008). Noteworthy is also that in more recent years, the consulting of eWOM is not always on a voluntary or intended basis; even if consumers aren’t actively looking for eWOM, viral effects ensure it also reaches them effortlessly (Gunawan & Huarng, 2015).

(13)

13

2.3 Online Consumer Reviews

(14)

14

This suspicion stems from early research on social cognition which confirmed that when making judgments, people tend to underuse base-rate information (e.g., aggregated ratings) and rely almost exclusively on individual information (e.g., consumer reviews) if both types of information are available (Borgida & Nisbett, 1977; Locksley, Hepburn & Ortiz, 1982). Because consumer reviews substitute or complement other sources of information present on the webshop which may suffer from marketing-based subjectivity biases, reviews are a valuable tool for consumers in their selection process (Bickard & Schindler, 2001). For this reason, this study focuses on exposure to individual consumer reviews as the antecedents of behavioural intentions. As online consumer reviews come in both positive- and negative forms with directionally different effects on consumer behaviour, both forms are included in this study.

(15)

15

2.4 The Information Adoption Model

(16)

16

Siegal, 2003; Park et al., 2007). If a review contains more understandable and objective comments with sufficient reasons of recommendation, it is more persuasive than a comment that expresses feelings and recommendations without specific reasons. Furthermore, because reviews posted on the internet are often anonymous, consumers generally do not easily accept or believe a review if it does not provide enough factual information in its argumentation (Ratchford, Talukdar & Lee, 2001). Because of previously presented arguments, a higher review quality implies a higher degree of information usefulness, and thus adaption.

(17)

17

2.5 The Relationship Between Perceived Review Quality and Behavioural Intentions

Perceived review quality is generally defined as ‘a consumer’s perception of the persuasive strength of argumentation used to convey an informational message’ (Bhattacherjee & Sanford, 2006). In this study, it is defined as the extent to which the readers of a review consider the arguments used in the review convincing in defending its position. Incorporating both argument quality and source credibility, it is best measured from the perspective of information characteristics in terms of credibility, understandability, sufficiency, and objectivity (Park, Lee & Han, 2007). On the other hand, behavioural intentions, as the outcome variable, is comprised of a social and economic aspect. Zeithaml, Berry & Parasuraman (1996) found that the use of a unidimensional measure of consumer behaviour would be troubling, because consumers react to information in multiple behavioural ways. Social intentions refer to cognitive reactions of consumers, such as further WOM intentions. Economic intentions on the other hand refer to a consumer’s financial reactions, manifested in purchase intention (Smith, Bolton & Wagner, 1999). Taken together, these behavioural intentions can be viewed as adequate indicators signalling whether consumers will engage with, remain with, or defect from a company (Zeithaml et al., 1996).

(18)

18

(Senecal & Nantel, 2004; Ruiz-Mafe, Chatzipanagiotou & Curras-Perez, 2018). By this reasoning, higher levels of perceived review quality are more likely to lead to a ‘change’ in consumer behaviour than lower levels. In addition, Vermeulen & Seegers (2009) show that whereas positive reviews improve consumers’ behavioural intentions, negative reviews result in a negative behavioural change.

If the perceived quality of a negative review is high, it is likely to be perceived as useful and to aid a consumer’s decision making journey, especially because a higher perceived quality will make the already diagnostic negative information even more valid for adoption. Under the logic of the information adoption model, a consumer is then likely to adopt the negative information, thus accepting the ‘warning’ given by the reviewer about avoiding behavioural intentions. Consequently, if negative online reviews serve to reveal potential threats and reduce the risks associated with the purchase of a product, they are expected to significantly reduce a consumer’s purchase and recommendation intentions (Purnawirawan, Eisend & De Pelsmacker, 2015). Consequently, the higher the perceived quality of a negative review, the higher the negative impact on behavioural intentions.

(19)

19

positive reviews with a higher perceived quality are expected to have a more positive effect on consumers’ behavioural intentions than reviews which are perceived to be of lower quality. This reasoning, in line with the information adoption model, results in the following hypothesis:

H1: A higher level of perceived review quality will (a) lower behavioural intentions when the product review is negative and (b) enhance behavioural intentions when the product review is positive

Aside the present unclarities in the effect of positive and negative eWOM messages, extant literature has not yet reached consensus as to why behavioural intentions vary largely from consumer to consumer individually (Gottschalk & Mafael, 2017). The area suffers from a lack of focus on how personal differences between individual eWOM receivers affect their behavioural intentions in different ways. Especially little attention has been given to how eWOM receivers might be attached to the brands they receive eWOM about, and for this reason brand commitment has often been suggested as an important moderating factor in eWOM studies (Wilson et al., 2017).

2.6 Brand Commitment

(20)

20

relationship with a particular brand (Moorman, Zaltman, & Deshpande, 1992). In essence, brand commitment represents a psychological attachment that provokes the innately human tendency to create stability among already acquired beliefs (Chaiken & Maheswaran, 1994). Hence, brand commitment reflects a strong emotional and motivational tie with a brand (Kim et al., 2014). The strength of a brand is known to affect the relationship between eWOM and sales, but brand managers increasingly realize that following online consumer conversations and monitoring opinionated posts is essential for the company. More and more firms incorporate social content in their marketing strategies and try to engage with consumer-led conversations (Deloitte, 2018). Early research found that knowing and liking a brand can change the way consumers interpret brand-related information (Ahluwalia, Burnkrant & Unnava, 2000; Ahluwalia, 2002). Furthermore, the level of brand commitment has been suggested to moderate a consumer’s behavioural responses to both negatively and positively valenced information regarding that brand (Funk & James, 2006). Consequently, brand commitment is expected to serve as a potent moderator in this study in order to uncover why eWOM behavioural effects differ per individual consumer.

2.7 The Moderation Effect of Brand Commitment

(21)

21

For negative information, psychological attachment to a brand leads to discounting of negative messages, or even complete rejection, when the messages are not consistent with a consumer’s attitudes or opinions (Crosby & Taylor, 1983). In other words, consumers read and evaluate negative information through a ‘biased lens’. Prior commitment and positive feelings about a brand have been found to attenuate the negativity effect of eWOM by preventing the adoption of negative messages in the first place (Kirmani, Sood and Bridges, 1999; Ahluwalia, 2000; Roehm and Brady, 2007). Hence, the impact of eWOM is strongly related to whether that eWOM is about the respondent’s preferred brand (East et al., 2008). Wilson et al. (2017) even found that highly committed consumers process negative eWOM defensively as the committed consumer feels the urge to prove his point and ‘defend’ the brand. Consumer are thought to self-identify with a brand, thus feeling like being personally attacked by negative reviews of other consumers. The authors suggest it would be interesting to see how the perceived strength of an argument influences this relationship (Wilson et al., 2017). In accordance, by drawing on the information adoption model and previous research this study predicts that commitment to a brand weakens the negative relationship between the perceived quality of negative eWOM information and a consumer’s behavioural intentions. This results in the following hypothesis:

H2: Brand commitment positively moderates the negative relationship between perceived review quality and behavioural intentions for negative reviews

(22)

22

perceived usefulness (Xia & Bechwati, 2008). When consumers perceive the information they receive as consistent with their prior knowledge or expectations, they have more confidence in adopting the information and using it in their decision journey (Sussman & Siegal, 2003). Consequently, it is expected that after exposure to positive reviews, brand commitment strengthens the relationship between perceived review quality and behavioural intentions. This argumentation results in the following hypothesis:

H3: Brand commitment positively moderates the positive relationship between perceived review quality and behavioural intentions for positive reviews.

(23)

23

likely to already poses high behavioural intentions before exposure to positive eWOM messages. This implies that for cases of positive eWOM there is less room for positive moderation to take place than there is for cases of negative eWOM. Following, the final hypothesis of this study proposes that:

H4: The moderation effect of brand commitment is stronger for negatively valenced reviews than for positively valenced reviews

2.8 Conceptual Model

The effects of the interrelating variables are presented in the conceptual model below, together with hypotheses 1 through 3.

(24)

24

3. METHODOLOGY

3.1 Research Design

(25)

25

Finally, the information adoption model suggests that consumers adopt information and make resulting decisions only if they also perceive the quality of said information to be high (Sussman & Siegal, 2003). This means that measuring behaviour after mere experimental manipulation by the researcher likely decreases the external validity of the research. Consequently, measuring the perceived quality of reviews in the manipulated condition group, while holding dependent variables constant, allows for the best causal exploration of all the hypothesized relationships.

3.2 Industry and Focal Brand Selection

(26)

26

3.3 Manipulation Conditions

To operationalize the 2x2 between-subjects experimental design, four separate conditions were created according to the variables under investigation. One condition for a positive online review with a high quality, one condition for a positive review with a low quality, one condition for a negative review with a high quality and finally, one condition for a negative review with a low quality.The following sub-section will elaborate on the development of the four different experiment conditions displayed in table 3.1.

Table 3.1: Overview of the Experiment Conditions

Condition Review Valence Review Quality

A B C D Positive Positive Negative Negative High Low High Low 3.4 Development of Stimuli

To make sure the experiment reflected reality as much as possible, the decision was made to adapt real-life reviews as stimuli without further manual manipulation. Adapting stimuli in such a manner offers the greatest validity and reliability, but requires careful and thorough selection of the adapted material (Kirk, 2013).

(27)

27

of the research. This left Amazon and eBay as potential source platforms, with Amazon chosen as the final online review source because of two reasons. First, the product page of a smartphone on Amazon.com is one of the first links a consumer is exposed to when he or she searches for a review in a search engine like Google. Second, eBay concerns itself with second-hand products while Amazon’s primary business is new products, and eWOM is known to be most impactful when consumers shop for new search products like is the case in our research (Tsao & Hsieh, 2015).

For the next step, the criteria of relevance, understandability, credibility and sufficiency were chosen from the perspective of information characteristics in order to carefully filter and select high- and low quality reviews. These characteristics are the embodiment of the scale items that measure the variable of ‘Perceived Review Quality’, and as a result were adapted as the relevant criteria for stimuli development. High-quality reviews are defined to be “product-relevant, understandable, and persuasive, with sufficient reasons based on facts about the product” while low-quality reviews are “emotional, subjective, and vacuous, with no information except expressions of subjective feelings or simple interjections” (Park et. al, 2007).

(28)

28

stimuli. It was decided that the two high quality reviews (A and C) needed at least a mean score of 5.5 (out of a 7-point Likert scale) on the fifth item: ‘In general, the quality of the review is high’. Furthermore, all other items could not score below a mean score of 4.5 for these reviews. For the low quality reviews (B and D), the maximum mean score for the fifth item was determined at 2.5, with the rest of the items restricted at a maximum mean of 3.5.For three out of four conditions (A, B and C) this yielded acceptably representative stimuli. However, for condition D the score on the fifth item of both tested stimuli was under 5 (4.8 and 4.2 respectively). Consequently, both potential stimuli were rejected as stimuli and a new, better suited stimulus for condition D had to be found.

After extensive search, a new review that adhered to the defined criteria was found on a different product page. This review was also related to the iPhone X, but with a different amount of Gigabytes of storage space. However, the review did not contain any references to the storage space and was therefore deemed satisfactory as a potential new stimulus. In order to find out if this new stimulus was more representative of its conditions than the 2 rejected stimuli, it was again tested in Qualtrics using the same measures as before (N = 10). This time, the stimulus was found to be sufficiently representative of a High Quality Negative Review, with the fifth item yielding a score of 5.90. All 9 investigated reviews and the statistics that determined either their acceptation or rejection of them can be found in Appendix A.

(29)

29

3.5 Measurements

In order to ensure reliability, all the measures for the relevant variables were derived from highly ranking business- and marketing journals. All individual scale items were measured on a seven point Likert scale, ranging from ‘strongly disagree’ to ‘strongly agree’. Although some researchers adopt a five-point Likert scale in order to draw comparisons with other research using five-point Likert scales (Saleh & Ryan, 1991), seven-point Likert scales are generally considered to be optimal in most instances of survey research and are especially reliable in the electronic distribution of experience metrics (Marsden & Wright, 2010; Finstad, 2010).

3.5.1 Main Constructs

The dependent variable of behavioural intentions was measured using items adapted from Zeithaml, Berry and Parasuraman (1996), which have been widely used in consumer behavioural- and marketing research. For the independent variable, perceived review quality was measured using items adopted from Park, Lee and Han (2007), which the authors based on the perspective of psychological information processing criteria (Galbraith, 1974). Finally, the moderator of brand commitment was measured using a three-item construct devised by Beatty, Kahle and Homer (1988). This tried and true measure has proven itself to be a very reliable measure of brand commitment in the field of WOM studies (Ahluwalia et al., 2000; Chang & Wu, 2014). The exact details of all main construct measures can be found in appendix B. 3.5.2 Control Variables

(30)

30

the relationship between eWOM adoption and behavioural intentions and therefore, gender is also as second control variable within this study (Wahid, 2007). Third, one important factor that is inherently characteristic to any communication through internet is culture; the internet removes the physical barriers to communication between consumers from different cultures, as is also the case with eWOM (De Mooij & Hofstede, 2002). Previous research has found that the impact of negative eWOM on consumer behaviour may be tempered for consumers hailing from countries with a higher score on Hofstede’s individualism metric (Goodrich & De Mooij, 2014). For this reason, this study employs an individualism control variable to ensure that the known effects of cultural differences are not excluded from the results. Including this control variable was accomplished by adopting the latest Hofstede scores on individualism for each of the respondents’ countries, leading to an ordinal arrangement of the nationalities present based on these scores from low to high. A list of these countries with their respective scores and re-arrangement can be found in appendix C.

(31)

31

Table 3.2: Overview of The Variables Used in This Study

Question Variable Type Number of items Example

1.1 – 1.4 Demographics Control 4 Gender, Age, Nationality, Education

2.1 – 2.3 General Attitude Toward Online Reviews

(Park, Lee & Han, 2007)

Covariate 3 ‘If I do not read online consumer reviews prior to purchase, I will feel worried about my decision’

3.1 – 3.3 Brand Commitment (Beatty & Kahle, 1988)

Moderator 3 ‘When another brand is on sale, I will still rather buy an Iphone.’ 4.1 – 4.5 Review Quality

(Park, Lee & Han, 2007)

Independent 5 ‘The review contained sufficient reasons supporting the opinions’ 5.1 – 5.3 Behavioural Intentions

(Zeithaml, Berry & Parasuraman, 1996)

Dependent 3 ‘I would tell my friends (looking to buy a new phone) to consider the brand I read about”

3.6 Data Collection Methods and Sampling

(32)

32

answer as truthfully as possible, and from their own personal point of view to reduce potential method biases. To further motivate the provision of valid and truthful data, the aim of this research was communicated in the introduction of the survey to the participants to stimulate interest and commitment to the research. Theory on market research also suggests to reduce the likelihood of incomplete data by not asking any sensitive questions (e.g. related to religion or personal income) and by explicitly naming the anonymity of the respondent and confidentiality of the provided answers in the survey (Malhotra & Peterson, 2014).

(33)

33

Finally, chance of random sampling errors decreases when the sample size is appropriately large enough for the respective research (Fowler, 2013). In order to define the ideal survey sample size, Schreiber et al., (2006) suggest that each parameter on the employed measure should have at least 10 participants. With a total of 18 parameters, ideal sample size is 180. This is validated by basing the sample size on a 95% confidence level and 5% margin of error where it is also deemed appropriate, especially when the population constitutes millions of consumers (Sekeran & Bougie, 2016). Hair, Tatham, Anderson and Black (2009) even regard five respondents per variable as the lower limit. However, this research involved different conditional groups, thus increasing the need for a sizeable sample. In addition, the total sample was later split for methodological purposes, which increases the preference for a larger sample.

3.7 Ethical Considerations

(34)

34

4. RESULTS

4.1 Data cleaning

The survey experiment collected sample data from October 18th until October 28th 2018 and yielded a total of 351 responses. A total of 78 respondents did not complete the questionnaire up to the final measurement question and were thus removed, case-wise, from the data set. Participants that had only failed to answer the question regarding participation in the giftcard draw were regarded as unfinished responses by Qualtrics, but were nevertheless included in the dataset. Furthermore, to control for random clicking all responses under 60 seconds were removed as suggested by Fowler (2013) which added up to a total of 100 removals. The final total sample amounted to N = 251. Moreover, all participants were randomly exposed to one of four stimuli in the survey experiment resulting in cell sizes A=59, B=68, C=60 and D=64.

4.2 Sample characteristics

(35)

35

Table 4.1: Total Socio-Demographic Sample Characteristics

Next, in order to form two sub-samples for later methodological purposes, an overview was necessary to check if each distinct experimental condition cell had roughly equal distributions. If such equal distribution is the case, adding two cells together (or splitting the total sample in two) to form a new sub-sample is allowed (Lin, 2013). Table 4.2 details the sample characteristics per condition group, which show that the cells are indeed fairly equal in their distributions.

Category N

251

%

100%

Total Sample Size

Gender Male Female 121 130 48,2% 51,8%

Age Group 18 and under

19-28 29-38 39-48 49-58 59 and over 6 178 7 16 28 16 2,5% 70,9% 2,8% 6,4% 11,2% 6,4%

Education Lower General Secondary Education

(36)

36

(37)

37

4.3 Factor Analysis

A confirmatory factor analysis was conducted to verify the validity of the adopted measurement scales and to ensure a good construct reliability. As the extraction method, a principal component analysis with forced factor extraction was employed. First, the independent variables ‘perceived review quality’ and the covariate ‘general attitude towards reviews’ were analysed. The aim of the factor analysis is to make sure that all items have high loadings on their respective factors. Hence, the following criteria were applied: each item must have a loading of 0.4 or higher into the correct factor, and each item cannot have a cross loading of greater than 0.4. One item of perceived review quality (PRQ3) loaded on the general attitude construct. This item consisted of the statement ‘The review is credible’ and the failure of loading on the correct dimension is most likely due to the finding in previous research that credibility is hard to determine in online reviews, as also discussed in the theoretical background (Verhagen et al., 2013). After removing said item, the independent measures in the analysis loaded highly on their respective components.

(38)

38

Table 4.3: Factor Loadings from the Factor Analysis Measure No. Perceived

Review Quality General Attitude Towards Reviews PRQ1 ,876 -,004 PRQ2 PRQ3 (Deleted) ,852 - ,035 - PRQ4 ,823 -,072 PRQ5 ,861 ,038 GA1 -,009 ,875 GA2 -,029 ,868 GA3 ,030 ,614

Measures that loaded on the same factor are indicated in bold

Table 4.4: Factor Loadings from the Factor Analysis of Brand Commitment

Measure no. Brand Commitment

BC1 0,899

BC2 0,949

BC3 0,945

Table 4.5: Factor Loadings from the Factor Analysis of Behavioural Intentions

Measure no. Behavioural Intentions

BI1 0,942

BI2 0,931

BI3 0,962

4.4 Descriptive Statistics, Construct Reliability and Correlations

(39)

39

Table 4.6: Descriptive Statistics, Cronbach’s Alpha and Correlations

** Significant at p < 0.01

4.5 Manipulation Check

In order to assess the successfulness of the manipulations in the experiment, a one-way ANOVA was conducted with the experimental conditions as the independent variable and perceived review quality as the dependent variable. In addition, a Levene’s test for equality of variances (p=0.028*) was performed to confirm that there was homogeneity of variances. The result showed that subjects in the high-quality-review conditions indeed had higher perceptions

Descriptive Statistics Reliability Correlations

(40)

40

of review quality than subjects in the low-quality-review conditions. More specifically, the perceived review quality decreased from condition A (M = 5.2 ±1.1), to condition B (M = 3,5 ±1.1) for the positive reviews, and from condition C (M = 5.4 ±0,9) to condition D (M = 2,1 ±0,71) for the negative reviews. The ANOVA revealed that the independent variable perceived review quality was significantly different for the different conditions presented to the survey respondents, F(3, 247)=169,971, p < .001. Because the raw data in this study is not significantly normally distributed but its residuals are, a Kruskal-Wallis test was conducted to double-check the findings of the ANOVA through analysis of the group medians. This test confirmed the statistically significant difference between the four conditions (see table 4.8). Therefore, the high- and low quality groups could be synthesized with their negative- or positive counterparts into two individual subsets. This sample split allows for the analysis of independent variable perceived review quality and moderator brand commitment in multiple regression analysis. Table 4.7 provides an overview of the manipulation analysis and the corresponding Levene test.

Table 4.7: Overview of the Manipulation Analysis and Levene Test

*Significant at the 0.05 level (2-tailed) DV= PRQ

** Significant at the 0.01 level (2-tailed)

CONDITION N Mean SD Levene Statistic ANOVA F-value

(41)

41

4.6 Requirements for Further Analysis

(42)

42

Table 4.8: Overview of manipulation tests

*Significant at the 0.05 level (2-tailed) DV= BI

** Significant at the 0.01 level (2-tailed)

Accordingly, the total sample (n=251) was split based on exposure to negatively (n=124) and positively (n=127) valenced reviews. To facilitate the comparison between the positive and negative review groups, we report the results from both sub-samples side by side in each following section. Before actual analysis by means of regression is possible, the data needs to be normally distributed. In linear regression, a frequent misconception is that the outcome has to be normally distributed, while the assumption is actually that the residuals have to be normally distributed (Greene, 2003). The P-P plot compares the observed cumulative distribution function (CDF) of the standardized residual to the expected CDF of the normal distribution. To confirm the normality of the residuals, the Q-Q plot compares the observed quantile with the theoretical quantile of a normal distribution. Both plots show that the standardized residuals follow an expected normal distribution, thus satisfying the assumption of normality required for multiple hierarchical regression in both sub-samples. Appendix E displays the P-P and Q-Q normality plots for both the positive and negative sub-samples.

Null Hypothesis Test Sig. Decision

The distribution of Perceived Review Quality is the same across the manipulated conditions

Independent Samples

Kruskal-Wallis Test

,00** Reject the null hypothesis

The distribution of Behavioural Intention is the same across categories of Valence Group

Independent Samples

Kruskal-Wallis Test

(43)

43

4.7 Regression Models

Hierarchical multiple regression analyses have been performed to test the five proposed hypotheses. First, the regression models assess whether perceived review quality has an influence on behavioural intentions in positively- and negatively valenced reviews. Subsequently, the moderating impact of brand commitment on these relationships is assessed. Finally, the relative strength of the moderation effect is compared between the positive and negative sub-samples. In addition, the conducted hierarchical multiple regression analyses are controlled for the influences of gender, age, general attitude towards reviews and individualism. Two independent analyses were conducted for the split sample, with each sub-sample assessed in three models, resulting in a total of six models. Table 4.9 provides an overview of all the hierarchical multiple regression results for behavioural intentions.

(44)

44

Table 4.9: Hierarchical Multiple Regression Results for Behavioural Intentions

Model 1 Model 2 Model 3

Unstandardized Coefficients Standard Error (VIF) Unstandardized Coefficients Standard Error (VIF) Unstandardized Coefficients Standard Error (VIF) Controls Gender Negative Positive -.077 .481 .346 (1.071) .311 (1.054) -.128 .536 .329 (1.073) .284 (1.056) -.108 .657* .322 (1.074) .278 (1.074) Age Negative Positive .012 -.010 .014 (1.110) .011 (1.099) .015 -.012 .013 (1.115) .010 (1.100) .018 -.009 .013 (1.129) .010 (1.074) General Att. Positive Negative .217 -.273 .168 (1.054) .160 (1.093) .241 -.267 .160 (1.056) .148 (1.079) .159 -.307* .161 (1.108) .142 (1.101) Individualism Negative Positive .378** -.263 .102 (1.007) .255 (1.078) .343** -.230 .097 (1.017) .235 (1.079) .309** -.239 .096 (1.041) .226 (1.079) Main effects PRQ Negative Positive -.607** .656** .160 (1.017) .140 (1.004) .-628** .693** .158 (1.020) .134 (1.010) Moderation effects PRQ x BC Negative Positive .375* -.458** .159 (1.103) .135 (1.048)

Model Summaries Model 1 Model 2 Model 3 F value Negative Positive 3.769* 1.721 .112 .053 .083 .022 6.224** 6.042** .209 .200 .175 .167 6.312** 7.376** .245 .269 .206 .233 R2 Negative Positive Adjusted R2 Negative Positive

Dependent Variable: Behavioural Intentions; Total N = 251, N negative= 124, N positive = 127 * Significant at p ≤0,05

(45)

45

4.8 Regression Findings

An examination of the regression results allows for empirical inferences regarding the acceptation or rejection of the proposed hypotheses, but also enables the discovery of other significant effects. First, it is noteworthy that three control variables have statistically significant effects in the third model. For the negative review sub-sample, the individualism score of the participant’s home country had a significant effect on the dependent variable behavioural intentions (B= 0.309, p <0.01). For the positive review sub-sample, gender (B=.657, p<0.01) and general attitude towards reviews (B= -0.307, p <0.05) were found to have significant effects on behavioural intentions. These control variables were only significant in the third (and final) model. The control variable age did not show a significant effect on the dependent variable in any of the models proposed in either of the sub-samples.

(46)

46

Third, the moderation effects of brand commitment on the main interaction effects were assessed. These moderation effects were only present in model 3 of both sub-samples, and their effects were significant in both cases. In the negative review sub-sample, brand commitment was found to have a significant positive moderation effect, B= 0.375, p < 0.05, on the negative relationship between perceived review quality and behavioural intentions, thus weakening that relationship with higher values of brand commitment. This confirms hypothesis 2, meaning that brand committed consumers are less sensitive to the negative impacts of negative reviews with a high perceived quality. Interestingly, even though the moderation effect of brand commitment was also significant for the positive review sub-set, its effects were found to have a negative effect (B= -0.458, p < 0.01) which was a conversely directional effect from that proposed in the corresponding hypothesis. Consequently, hypothesis 3 was rejected.

(47)

47

4.9 Hypotheses Testing Results Overview

Table 4.10 provides an overview of hypotheses testing results. Overall, concluded can be that support was found for hypotheses 1a, 1b and 2, rendering them accepted, while no support was found for hypotheses 3 and 4, rendering them rejected.

Table 4.10: Overview of Hypotheses Testing Results

Proposed Hypotheses Result

H1a: A higher level of perceived review quality will lower behavioural intentions when the product review is negative

Accepted

H1b: A higher level of perceived review quality enhance behavioural intentions when the product review is positive

Accepted

H2: Brand commitment positively moderates the negative relationship between perceived review quality and behavioural intentions for negative reviews

Accepted

H3: Brand commitment positively moderates the positive relationship between perceived review quality and behavioural intentions for positive reviews.

H4: The moderation effect of brand commitment is stronger for negatively valenced reviews than for positively valenced reviews.

Rejected

(48)

48

5. DISCUSSION

5.1 Restating the Research Aim

This research about electronic word-of-mouth examined the effects of perceived quality of both positive and negative eWOM on consumers’ behavioural intentions (e.g., purchase intentions, recommendations) by focussing on online consumers reviews. In addition, this research aimed to uncover the moderation effects of a brand commitment on the examined relationships to provide empirical evidence and an explanation for individual behavioural differences between consumers.

5.2 Perceived Review Quality and Behavioural Intentions

(49)

49

intentions than positive information was expected to be beneficial (Chevalier & Mayzlin, 2006). Nevertheless, positive reviews with a high perceived qualities were found to have a stronger impact on behavioural intentions than their high quality negative counterparts. A possible explanation for this finding is offered by Zhang et al. (2010), who found that for consumer goods associated with promotion goals(i.e. reaching a ‘positive end state’) such as mobile phones, positive reviews are more persuasive than negative reviews. The reverse is true for prevention goal products (preventing a ‘negative end state’) such as antivirus software or insurance. These findings imply that it is important to be more nuanced in generalizing the negativity bias across all product types in eWOM research, as high quality positive reviews might have a bigger impact than is commonly thought for specific product types.

5.3 The Moderation Effect of Brand Commitment

(50)

50

reviews. This results in counter-arguing or ‘fighting back’, thereby alleviating negative behavioural effects by strengthening feelings towards the brand (Wilson et al., 2017).

(51)

51

diagnostically evaluate the quality of the reviews, and thus the perceived quality of a review has a bigger effect on their behavioural intentions.

5.4 Research Implications 5.4.1 Academic

This study has contributed to the expanding field of eWOM literature by uncovering why recent research has been largely inconclusive regarding the behavioural effects of positive and negative eWOM. Previous research has focused on comparing the direct effects of exposure to positive or negative eWOM, restricting itself to manipulation (Verhagen et al., 2013). Consequently, although influential studies in the field have proven that eWOM affects consumer behaviour, explorations of the underlying mechanics of these effects were largely absent (King et al., 2014). Because the concept of quality is highly subjective, it was expected to be a critical factor in determining the differences in effect of online reviews for different individuals. By moving beyond mere manipulation of quality towards subjective measurement, this study was able to explain the underlying mechanisms of behavioural eWOM effects. To the best of the author’s knowledge, this study is thereby the first to provide empirical evidence for the impact of perceived quality of both positive- and negative eWOM messages on behavioural intentions. Future eWOM research would benefit from trying to also steer clear of researcher-subjective definitions of quality.

(52)

52 5.4.2 Managerial

From a managerial perspective, this study provides marketers with a better frame of reference to understand the influence of eWOM on consumer behaviour. Regarding negative eWOM, the finding that a high perceived quality leads to decreased behavioural intentions should motivate managers to resolve issues quickly, as overlooking consumer publications could have harmful effects on sales (Erkan & Evans, 2016). Detailed negative reviews have the highest chance of causing damage to consumer behaviour, but also offer the most detail to what the poster of the message is unhappy about. Quick recovery strategies have been known to reduce the damage done by reviews if the poster of the message deletes the message after the issue has been resolved, or counter-balances the negative effect with a positive comment about the timely resolving of the issue (Cambra-Fierro, Melero & Sese, 2015). Taking negative opinions into account can hence make the customer feel appreciated, leading to increased satisfaction and decreased future negative e-WOM. Furthermore, negative opinions on the web could reveal valuable insights about directions for improvement of the product. Moreover, this study’s confirmation regarding the moderation effect of brand commitment after negative review exposure indicates that companies must try to increase consumer brand commitment in order to reduce the aftermath of negative reviews.

(53)

53

Finally, the novel finding that a higher brand commitment negatively moderates the positive impact of perceived review quality on behavioural intentions or positive reviews offers the suggestion that marketing managers should focus on exposing positive reviews to non-committed consumers more than to non-committed consumers. From a marketing perspective, a valuable suggestion would be to seek out platforms where non-committed consumers are present, instead of relying on brand-communities and groups of already committed consumers. In conclusion, managers should consider strategies that generate high levels of brand commitment to counter negative eWOM effects, but should also focus on non-customers because brand commitment is found to decrease the marginal positive effects of eWOM due to redundant confirmation of existing beliefs.

5.5 Limitations and future research avenues.

The results presented in this study are subject to a number of limitations, and should be interpreted with those limitations in mind. However, these limitations also open op new avenues for future research.

(54)

54

replicated using probability sampling or is conducted in two different countries and results are then compared.

Second, the chosen industry and single focal brand may also influence the generalizability of this research. In this research only the smartphone industry was chosen as a research context, which offers a product with a relatively high switching cost. However, previous studies have indicated that eWOM about experiential and search goods may offer different behavioural outcomes (Xia & Bechwati, 2008). The generalizability of the research would benefit if a similar study would be performed with multiple product types and brands.

(55)

55

6. CONCLUSION

Consumers increasingly use customer-driven platforms on webshops to inform themselves of opinions about products they consider to buy. However, such reviews may be positive and negative and their content quality greatly differs. Both of these characteristics have a major influence on the strength of a consumer’s behavioural intentions. In addition, how much a consumer is committed to a brand can greatly colour the information adoption of such reviews, thus further altering the resulting behavioural intentions. By analysing these factors, this research set out to uncover the mechanisms underlying behavioural intentions of consumers after exposure to positive and negative reviews.

(56)

56

(57)

57

7. REFERENCES

Ahluwalia, R., Burnkrant, R. and Unnava, H. (2000) ‘Consumer Response to Negative Publicity: The Moderating Role of Commitment’, Journal of Marketing Research, 37(2), pp. 203-214.

Ahluwalia, R. (2002) ‘How Prevalent Is the Negativity Effect in Consumer Environments?’, Journal of Consumer Research, 29(2), pp. 270-279.

Amblee, N. and Bui, T. (2011) ‘Harnessing the Influence of Social Proof in Online Shopping: The Effect of Electronic Word of Mouth on Sales of Digital Microproducts’,

International Journal of Electronic Commerce, 16(2), pp .91-114.

Andersson, U. Cuervo-Cazurra, A. and Nielsen B. B. (2014) From the Editors: Explaining interaction effects within and across levels of analysis. Journal of International Business Studies, 45(9), 1063-1071.

Arndt, J. (1967) ‘Role of product-Related conversations in the diffusion of a new product’, Journal of Marketing Research, 4(3), pp. 291-291.

Bailey, J. E., and Pearson, S. W. (1983) ‘Development of a tool for measuring and analyzing computer user satisfaction’, Management Science, 29(5), pp. 530-545.

Balaji, M. S., Khong, K. W., and Chong, A. Y. L. (2016) ‘Determinants of negative word-of-mouth communication using social networking sites’, Information &

Management, 53(4), pp. 528-540.

Bambauer-Sachse, S., and Mangold, S. (2011) ‘Brand equity dilution through negative online word-of-mouth communication’, Journal of Retailing and Consumer

Services, 18(1), pp. 38-45.

Bansal, H. S., and Voyer, P. A. (2000) ‘Word-of-mouth processes within a services purchase decision context’, Journal of Service Research, 3(2), pp. 166-177.

Berger, J. (2014) ‘Word of mouth and interpersonal communication: A review and directions for future research’, Journal of Consumer Psychology, 24(4), pp. 586-607.

Bhattacherjee, A. and Sanford, C. (2006) ‘Influence processes for information technology acceptance: An elaboration likelihood model’, MIS Quarterly, pp. 805-825.

Bickart, B. and Schindler, R. M. (2001) ‘Internet forums as influential sources of consumer information’, Journal of Interactive Marketing, 15(3), pp. 31-40

(58)

58

Bougie, R., Pieters, R. and Zeelenberg, M. (2003) ‘Angry customers don't come back, they get back: The experience and behavioral implications of anger and dissatisfaction in services’, Journal of the Academy of Marketing Science, 31(4), pp. 377-393.

Cambra-Fierro, J., Melero, I. and Sese, F. J. (2015) ‘Managing complaints to improve customer profitability’, Journal of Retailing, 91(1), pp. 109-124.

Chaiken, S. and Eagly, A. H. (1976) ‘Communication modality as a determinant of message persuasiveness and message comprehensibility’, Journal of Personality and Social Psychology, 34(4), pp. 605.

Chaiken, S. and Maheswaran, D. (1994) ‘Heuristic processing can bias systematic processing: effects of source credibility, argument ambiguity, and task importance on attitude

judgment’, Journal of Personality and Social Psychology, 66(3), pp. 460.

Chang, H. H. and Wu, L. H. (2014) ‘An examination of negative e-WOM adoption: Brand commitment as a moderator’, Decision Support Systems, 59, pp. 206-218.

Chevalier, J. A. and Mayzlin, D. (2006) ‘The effect of word of mouth on sales: Online book reviews. Journal of Marketing Research, 43(3), 345-354.

Cheung, C. M., Lee, M. K. and Rabjohn, N. (2008) ‘The impact of electronic word-of-mouth: The adoption of online opinions in online customer communities’, Internet

Research, 18(3), pp. 229-247.

Cheung, M. Y., Luo, C., Sia, C. L. and Chen, H. (2009) ‘Credibility of electronic word-of-mouth: Informational and normative determinants of on-line consumer

recommendations’, International Journal of Electronic Commerce, 13(4), pp. 9-38. Christodoulides, G. and De Chernatony, L. (2010) ‘Consumer-based brand equity

conceptualization and measurement: A literature review’, International Journal of Market Research, 52(1), pp. 43-66.

Chong, A. Y. L., Khong, K. W., Ma, T., McCabe, S. and Wang, Y. (2018) ‘Analyzing key influences of tourists’ acceptance of online reviews in travel decisions’, Internet Research, 28(3), pp. 564-586.

Çifci, S., Ekinci, Y., Whyatt, G., Japutra, A., Molinillo, S., & Siala, H. (2016). A cross validation of Consumer-Based Brand Equity models: Driving customer equity in retail brands. Journal of Business Research, 69(9), 3740-3747.

(59)

59

Crosby, L. A. and Taylor, J. R. (1983). ‘Psychological commitment and its effects on post-decision evaluation and preference stability among voters’, Journal of Consumer Research, 9(4), pp. 413-431.

Davis, F. D. (1989) ‘Perceived usefulness, perceived ease of use, and user acceptance of information technology’, MIS Quarterly, 13(3), pp. 319-340.

Deloitte (2018) ‘The Deloitte Consumer Review: Digital Predictions 2018, A Deloitte Insight Report’. Available at

https://www2.deloitte.com/content/dam/Deloitte/uk/Documents/consumer-business/deloitte-uk-consumer-review-digital-predictions-2018.pdf).

De Mooij, M., & Hofstede, G. (2002) ‘Convergence and divergence in consumer behavior: implications for international retailing’, Journal of Retailing, 78(1), pp. 61-69. Duan, W., Gu, B. and Whinston, A. B. (2008) ‘The dynamics of online word-of-mouth and

product sales—An empirical investigation of the movie industry’, Journal of Retailing, 84(2), pp. 233-242.

East, R., Hammond, K. and Lomax, W. (2008) ‘Measuring the impact of positive and negative word of mouth on brand purchase probability’, International Journal of Research in Marketing, 25(3), pp. 215-224.

Erkan, I. and Evans, C. (2016) ‘The influence of eWOM in social media on consumers’ purchase intentions: An extended approach to information adoption’, Computers in Human Behavior, 61, pp. 47-55.

Fan, W. and Gordon, M. D. (2014) ‘The Power of Social Media Analytics’, Communications of the ACM, Vol. 57 No. 6, 74-81.

Filieri, R. (2015) ‘What makes online reviews helpful? A diagnosticity-adoption framework to explain informational and normative influences in e-WOM’, Journal of

Business Research, 68(6), pp. 1261-1270.

Finstad, K. (2010) ‘The usability metric for user experience’, Interacting with Computers, 22(5), pp. 323-327.

Fournier, S. (1998) ‘Consumers and their brands: Developing relationship theory in consumer research’, Journal of Consumer Research, 24(4), 343-373.

(60)

60

Funk, D. C. and James, J. D. (2006) ‘Consumer loyalty: The meaning of attachment in the development of sport team allegiance’, Journal of Sport Management, 20(2), pp. 189-217.

Galbraith, J. R. (1974), ‘Organization design: An information processing view’, Interfaces, 4(3), pp. 28-36.

Godes, D. and Mayzlin, D (2004), ‘Using online conversations to study word-of-mouth communication’, Marketing Science, 23(4), pp. 545-560.

Goodrich, K. and De Mooij, M. (2014) ‘How ‘social’ are social media? A cross-cultural comparison of online and offline purchase decision influences’, Journal of Marketing Communications, 20(1-2), pp. 103-116.

Gottschalk, S. A. and Mafael, A. (2017) ‘Cutting through the online review jungle—

Investigating selective eWOM processing’, Journal of Interactive Marketing, 37, pp. 89-104.

Greene, W. H. (2003) Econometric analysis. Pearson Education India.

Greenfield, T. (2016) Research methods for postgraduates. John Wiley & Sons. Grewal, D., Monroe, K. B. and Krishnan, R. (1998) ‘The effects of price-comparison

advertising on buyers' perceptions of acquisition value, transaction value, and behavioral intentions’, The Journal of Marketing, pp. 46-59

Gunawan, D. D. and Huarng, K. H. (2015) ‘Viral effects of social network and media on consumers’ purchase intention’, Journal of Business Research, 68(11), pp. 2237-2241. Gupta, P. and Harris, J. (2010). ‘How e-WOM recommendations influence product

consideration and quality of choice: A motivation to process information perspective’, Journal of Business Research, 63(9-10), pp. 1041-1049.

Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E. and Tatham, R. L.

(2009) Multivariate Data Analysis. 7th edn. Upper Saddle River, NJ: Prentice hall. Hayes, A. F. (2017) Introduction to mediation, moderation, and conditional process

analysis: A regression-based approach. Guilford Publications.

Hennig-Thurau, T., Gwinner, K., Walsh, G. and Gremler, 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), pp. 38-52. Ho-Dac, N. N., Carson, S. J. and Moore, W. L. (2013) ‘The effects of positive and negative

(61)

61

Hofstede, G. (2018) Compare countries - Hofstede Insights. Available at:

https://www.hofstede-insights.com/product/compare-countries/ (Accessed 27 Oct. 2018).

Holbrook, M. B. (1993) ‘Nostalgia and consumption preferences: Some emerging patterns of consumer tastes’, Journal of Consumer research, 20(2), pp. 245-256.

Kahneman, D. and Tversky, A. (1984) ‘Choices, values, and frames’, American Psychologist, 39(4), pp. 341.

Kim, K., Park, J. and Kim, J. (2014) ‘Consumer–brand relationship quality: When and how it helps brand extensions’, Journal of Business Research, 67(4), pp. 591-597.

King, R. A., Racherla, P. and Bush, V. D. (2014) ‘What we know and don't know about online word-of-mouth: A review and synthesis of the literature’, Journal of Interactive Marketing, 28(3), pp. 167-183.

Kirk, R.E. (2013) Experimental design: Procedures for the Behavioral Sciences. 4th edn. Los Angeles, Calif.: Sage.

Kirmani, A., Sood, S. and Bridges, S. (1999) ‘The ownership effect in consumer responses to brand line stretches’, The Journal of Marketing, 63(1), pp. 88-101.

Kumar, N. and Benbasat, I. (2006) ‘The influence of recommendations and consumer reviews on evaluations of websites’, Information Systems Research, 17(4), pp. 425-439.

Laczniak, R. N., DeCarlo, T. E. and Ramaswami, S. N. (2001) ‘Consumers’ responses to negative word‐of‐mouth communication: An attribution theory perspective’, Journal of Consumer Psychology, 11(1), pp. 57-73.

Lee, K.-T. and Koo, D.-M. (2015) ‘Evaluating right versus just evaluating online consumer reviews’, Computers in Human Behavior, 45, pp. 316-327.

Lee, M. and Youn, S. (2009) ‘Electronic word of mouth (eWOM) How eWOM platforms influence consumer product judgement’, International Journal of Advertising, 28(3), pp. 473-499.

Lin, W. (2013) ‘Agnostic notes on regression adjustments to experimental data: Reexamining Freedman’s critique’, The Annals of Applied Statistics, 7(1), pp. 295-318.

Liu, Y. (2006) ‘Word of mouth for movies: Its dynamics and impact on box office revenue’, Journal of Marketing, 70(3), pp. 74-89.

(62)

62

López-López, I. and Parra, J. F. (2016), ‘Is a most helpful eWOM review really helpful? The impact of conflicting aggregate valence and consumer’s goals on product

attitude’, Internet Research, 26(4), pp. 827-844.

Malhotra, N. K. and Peterson, M. (2014) Basic marketing research. Pearson.

Matell, M. S. and Jacoby, J. (1971). ‘Is There an Optimal Number of Alternatives for Likert Scale Items? Study I: Reliability and Validity’, Educational and Psychological Measurement, pp. 657-674.

Marsden, P. V. and Wright, J. D. (2010) Handbook of Survey Research. Emerald Group Publishing.

Moorman , C., Zaltman, G. and Deshpande, R. (1992), ‘Relationships between providers and users of market research: The dynamics of trust within and between organizations’, Journal of Marketing Research, 29(3): pp. 314–328.

Nonaka, I. (1994) ‘A dynamic theory of organizational knowledge creation’, Organization Science, 5(1), pp. 14-37.

Nunnally, J. C. (1978) Psychometric theory. 2nd edn. McGraw-Hill. Hillsdale, NJ.

O'Cass, A. and Lim, K. (2002) ‘The influence of brand associations on brand preference and purchase intention: An Asian perspective on brand associations’, Journal of

International Consumer Marketing, 14(2-3), pp. 41-71.

Park, D. H., Lee, J. and Han, I. (2007). ‘The effect of on-line consumer reviews on consumer purchasing intention: The moderating role of involvement’ International Journal of Electronic Commerce, 11(4) pp. 125-148.

Park, D. H. and Kim, S. (2008). ‘The effects of consumer knowledge on message processing of electronic word-of-mouth via online consumer reviews’, Electronic Commerce Research and Applications, 7(4), pp. 399-410.

Peighambari, K., Sattari, S., Kordestani, A. and Oghazi, P. (2016) ‘Consumer Behavior Research: A Synthesis of the Recent Literature’, SAGE Open, 6(2), pp. 1-9. Petty, R.E. and Cacioppo, J.T. (1986) ‘The elaboration likelihood model of persuasion’

In Communication and Persuasion. Springer: New York, NY, pp. 1-24.

The Subjectively Objective Consumer: How Perceived Electronic Word-of-Mouth Quality and Brand Commitment Affect Behavioural Intentions