UNIVERSITY OF GRONINGEN
FACULTY OF ECONOMICS AND BUSINESS
Antecedents of Online Consumer Review Usage:
The Role of Shopping Orientation, Social Influence, and Product Type
Author: D.H.D. Wang
1
UNIVERSITY OF GRONINGEN
FACULTY OF ECONOMICS AND BUSINESS
MSc. Marketing
Master Thesis
Antecedents of Online Consumer Review Usage:
The Role of Shopping Orientation, Social Influence, and Product Type
2 Management Summary
The Internet has become a successful medium for the distribution of goods and services, and has grown rapidly ever since. Alongside the growth of the Internet, information exchange in the form of electronic word of mouth between consumers has gained popularity. Research on consumer review usage, however, is quite new and still limited. This study examined the antecedents of the usage of online consumer reviews and provides insights in drivers of consumer review usage. Existing literature about the drivers of consumer review usage is extended by modeling the consumer‟s shopping orientation, social influence, and product type against the usage frequency of consumer reviews, as these factors can have a significant impact on the (search) behavior.
By using a large-scale, Internet-based survey, data was collected through various channels. Prior to interpretation of the results, our data was tested extensively on inconsistencies and whether or not the residual assumptions of the models held. Our results showed that price consciousness actually has a positive impact, instead of the hypothesized negative effect. Informational influence also had a positive effect on the consumer review usage. The consumer‟s risk aversion, normative influence, and the moderating effects of product type did not have a significant impact. The main effect of product type, however, did have a significant impact on the usage frequency of consumers review; consumer review usage was higher for experience goods than search goods.
These results create insights in how the level of consumer review usage is influenced. Based on the customer characteristics and the types of products companies sell, managers can determine which customer groups will most likely engage in eWOM, and hence, use and value consumer reviews. This may help managers to decide whether the introduction of an online review system is useful for them or not, and how managers can optimize their current consumer review system. This is important as positive experience with consumer reviews and positive product ratings can, in turn, influence the consumer‟s purchasing behavior and decisions. Based on our findings, we suggest managers who are dealing with or attracting price conscious consumers to implement online consumer review systems. This also holds for companies that are dealing with experience goods.
3 Table of Contents 1. Introduction ... 4 2. Theoretical Framework ... 8 2.1 Shopping Orientation ... 8 2.2 Social Influence ... 10 2.3 Product Type ... 11 2.4 Control Variables ... 13 3. Research Methodology ... 14
3.1 Data Collection Method ... 15
3.2 Survey Development, Measurement, and Manipulation ... 15
3.3 Validity and Reliability of Measures ... 16
3.4 Model Specification and Method of Analysis ... 19
4. Results ... 19 4.1 Data ... 21 4.1.1 Sample Description... 21 4.1.2 Data Collection ... 22 4.1.3 Data Purification ... 22 4.2 Assumption Tests ... 23 4.2.1 Normality Test ... 23 4.2.2 Homoscedasticity Test ... 24 4.2.3 Collinearity Test ... 25 4.3 Model Selection... 25 4.4 Predictive Validity... 26 4.5 Hypothesis Testing ... 27 5. Discussion... 30 6. Management Implications ... 32
7. Limitations and Further Research ... 33
References ... 35
4 1. Introduction
In the course of years, the Internet has become a successful medium for both the distribution of goods and services, and the communication of product information by marketers (Rose and Samouel 2009). In 2009, Western Europe‟s online retail sales totaled $150.2 billion and are expected to increase in both sales and the number of online shoppers (Chen 2012). Shopping on the internet is a relatively new phenomenon; not until the late 1990s, online shopping has taken off (Zhou, Dai, and Zhang 2007). Various consumer factors contributed to the acceptance of online shopping, such as demographics, shopping orientation and motivation, internet experience, and online experience (for a detailed description, see Zhou, Dai, and Zhang 2007).
With the rapid diffusion of the Internet, information exchange in the form of electronic word of mouth (eWOM) between consumers has become ever frequent and easy (Floh, Koller, and Zauner 2013). Websites entirely dedicated to organizing recommendations and reviews of numerous products and services, such as Yelp.com, Epinions.com and TripAdvisor.com, have sprouted and gained popularity over the past decade (Racherla and Friske 2012; Robson et al. 2013). It is thought that consumer‟s purchasing decision can be significantly influenced by online consumer reviews as consumer reviews are a good proxy for overall word of mouth (Zhu and Zhang 2010). In fact, recent studies have shown that online consumer reviews can have a significant impact on product sales (Chevalier and Mayzlin 2006; Cui, Lui, and Guo 2012; Zhu and Zhang 2010). In some cases, consumer reviews are even being fraudulently manipulated to persuade the consumer in purchasing the „wrong‟ item by providing inauthentic information (Bambauer-Sachse and Mangold 2013; Hu et al. 2011; Jensen et al. 2013).
5
example, discussion boards for an extensive period of time (Breazeale 2009). Moreover, Hartman, Hunt, and Childers (2013) noted that eWOM tends to include both positive and negative information and consists of readily available information from multiple sources, organized for consumers. Because of the difference and the newness of eWOM compared to traditional WOM, it is crucial to examine eWOM in further detail. Meuter, McCabe, and Curran (2013) indeed found differences in the influence of WOM compared to eWOM.
As noted earlier, eWOM is tied closely to online consumer reviews. Examining the level in which consumers engage in eWOM, and thus, use consumer reviews, can be useful as consumer reviews can influence purchase decisions and have an effect on product sales. However, understanding the concept of consumer review usage is rather complex as many factors can influence the level in which consumers read information on these so-called consumer opinion platforms. Hennig-Thurau and Walsh (2003) identified five reading motives: to obtain buying-related information, to learn how a product is to be consumed, social orientation through information, community motive, and remuneration. Largely in line with these motives, Burton and Khammash (2010) distinguished seven themes why people read consumer-generated reviews: decision involvement, product involvement, economic involvement, consumer empowerment, self-involvement, social involvement, and site involvement. Reading consumer reviews is helpful in buying decision making, getting information about products, understanding the nature of the perceived uniqueness and quality of information, expanding the reader‟s general knowledge, and facilitating socialization. Posting consumer reviews, on the other hand, may be motivated by other reasons. Consumers may post reviews to gain self-approval or social approval and to express satisfaction or dissatisfaction (Chen, Fay, and Wang 2011). Hence, it is important to extend previous studies to get a better understanding in the drivers of online review usage.
Many studies that are conducted explain the factors that influence the effectiveness (i.e. on purchase intention, conversion rates, sales etc.) of consumer reviews. For example, Ludwig et al. (2013) examined the effects of the linguistic style of reviews, while others looked at the effects of product characteristics and consumer characteristics (Zhu and Zhang 2010), the volume and valence of product reviews (Cui, Lui, and Guo 2012), or the sequence in online reviews (Purnawirawan, De Pelsmacker, and Dens 2012).
6
reviews, and compared the differences between Korean and American consumers. Rose and Samouel (2009) modeled internal psychological and external market-driven determinants of the amount of online consumer information search such as perceived cost of online search, ability to search online, prior memory structure and the size of consideration set, while others examined the effects of, among others, perceived risk and perceived costs on consumers‟ online information search (Kulviwat, Guo, and Engchanil 2004; Maity, Hsu, and Pelton 2012). The antecedents of consumer review usage can, however, be explained by and extended with other factors as we illustrate in the next paragraphs.
First, “orientation is an interactive component controlling the tendency of behavior” (Kohijoki and Marjanen 2013, p. 166). Consumers with distinct shopping orientations have different shopping behaviors both in terms of frequency of internet purchases and the extent of information search about products or services while shopping online (Seock and Bailey 2008). Hence, shopping orientation can have an influence on the consumer‟s behavior. Researchers have investigated the concept of shopping orientation extensively from various perspectives, resulting in a wide variety of definitions of the concept (Mejri, Debabi, and Nasraoui 2012). Previous studies have shown that shopping orientation indeed has impact on different behavioral outcomes. For instance, Büttner, Florack, and Göritz (2013) examined cognitive procedures of shopping orientations, Arnold and Reynolds (2003) investigated motives for why consumers go shopping, and Rigopoulou, Tsiotsou, and Kehagias (2008) derived shopping orientations segments based on store-choice criteria and satisfaction. Others examined the effects of shopping orientations on the product usage rates (Darden and Reynolds 1971), online purchase intention (Brown, Pope, and Voges 2003), and the willingness to purchase clothes online (Hansen and Jensen 2009).
7
Reed II 1998), the decision to shop online (Lee et al. 2011), and online product choices (Huang and Chen 2006). Bonfield (1974) argued that the influences of social influences are situation bound and various situations may lead to other behavioral outcomes.
Third, from the early seventies on, the product type has been linked to the level of information search (Nelson 1970, 1974). Park and Lee (2009b) distinguished two product types: search and experience. Due to the different nature of products, whether it is a search or experience good, the way how consumers acquire and process information to make decisions differs, which may result in differences in online search behavior (Huang, Lurie, and Mitra 2009). Indeed, a study by Johnson et al. (2004) confirmed low information search volume for search goods and Park, Yoon, and Lee (2009) found that females use significant more customer reviews for experience goods than search goods. Therefore, in the online information search context, the product type should be taken into account.
8 2. Theoretical Framework
2.1 Shopping Orientation
9
Price consciousness. Seock and Bailey (2008) showed a significant positive relationship between a consumer‟s price consciousness and the degree of online information searches. This, however, does not necessarily imply that an increase in the degree of online information searches leads to an overall greater exposure to, hence usage of, online reviews. In fact, the authors argue that price conscious consumers are more likely to spend time searching for promotional deals and sales, and comparing prices of products offered by different companies. Consumer decision strategies can be processed through selective processing where only information of the most important factor/attribute is processed (Bettman, Luce, and Payne 1998). The choice process of highly price conscious consumers involves highly selective processing of information about price. In general, when consumer are more selective in information processing, other factors, such as the usage of consumer reviews, become less salient which may result in the omittance of such information. Based on these findings, we assume that the degree of the consumer‟s price consciousness impacts the usage frequency of online reviews negatively.
H1: The more price conscious a consumer, the lesser usage frequency of consumer reviews.
Degree of risk aversion. One of the benefits of viewing online consumer reviews is the ability to reduce (purchase) risk (Park and Lee 2009b). Risk can be defined as “reflecting variations in the distribution of possible outcomes, their likelihood and their subjective values” (Mitchell 1999, p. 167). In contrast with information generated by the manufacturer, personal opinions online are perceived as more trustworthy, and are therefore, perceived less risky (Burton and Khammash 2010). In the context of shopping orientations, the consumer‟s risk-consciousness can be seen as a part of the concept (Lee, Kim, and Lee 2013). A study conducted by Chaudhuri (2000) showed that the degree of perceived product risk has a significant impact on the level of information search. In offline contexts, Beatty and Smith (1987) suggested that risk aversion, the consumer‟s tendency to avoid losses (Grant, Clarke, and Kyriazis 2007), positively impacts the amount of search. This is also reflected in the financial sector where Tseng (2013) suggested that individuals with greater risk aversion increase information seeking. Based on these findings, we hypothesize:
10
2.2 Social Influence
In the fifties, Deutsch and Gerard (1955) distinguished two types of social influence that influence consumer behavior differently: normative and informational influence. The authors defined normative social influence as “an influence to conform with the positive expectations of another,” while informational social influence is defined as “an influence to accept information obtained from another as evidence about reality” (p. 629). Previous studies used the term “group influence” to describe social influences, instead of distinguishing different types of social influence. Consumers who post and/or view online reviews can be seen as members of a social group of online product reviewers, because an online product reviewer may influence the purchase decision of others, but in turn, is also being influenced by the ones who already have posted a product review (Sridhar 2012). Social influences result from human‟s nature to imitate in order to be accepted by others and to be safe (Huang and Chen 2006).
Normative influence. Normative social influence concerns conformance with the expectations of other persons or groups, of which the greatest influence is usually exerted within primary reference groups such as the immediate friends or family (Lord, Lee, and Choong 2001). Groups who share information online often face the issue of anonymity (Wodzicki et al. 2011). Dholakia, Bagozzi, and Pearo (2004) suggested that social influence is negatively related to the anonymity of members in a community as the people do not know each other and their motives are self-referent. Moreover, Gerard and Deutsch (1955) argued that normative influence depends on social pressure, which is greatest when group members are identifiable. Also, Latané (1981) found that the immediacy of its members, in terms of their space and proximity, affects social influence positively. These findings lead to the assumption that a consumer who is more prone to normative influences will make less use of consumer reviews because of the anonymous nature of the people who post reviews and because online consumers cannot be considered as one‟s immediate social relations due to the diversity of the internet users.
H3: The higher the degree of the consumer‟s susceptibility to normative influence, the lower the online review usage.
11
Dholakia, Basuroy, and Soltysinski (2002) suggest that in an online context, informational influence is expected to play a central role in influencing consumers because when making good decisions online, only informational motives are present. Therefore, we assume that the degree of the consumer‟s susceptibility to informational influence (i.e. in what extent consumers make decisions based on informational influences) will translate to a higher usage of online consumer reviews.
H4: The higher the degree of the consumer‟s susceptibility to informational influence, the higher the online review usage.
2.3 Product Type
12
Sinha and Batra (1999) noted that price conscious consumers may be reluctant to pay for a product‟s distinguishing features if the price difference for these features is too large. In this sense, trade-offs have to be made between higher price and potential benefits, such as an increase in quality. Product quality is harder to be determined for experience than for search goods as the product attributes for experience goods are only known after consumption and, therefore, the information search is likely to be higher for experience goods in order to determine the product quality. Based on this logic, we expect that the consumer‟s price consciousness has a greater (negative) effect on the usage level of consumer review for search goods than for experience goods.
The difficulty of determining product attributes for experience goods can also have an impact on the relationship between the consumer‟s degree of risk aversion and susceptibility to informational influence on the consumer review usage. In order to minimize, for example, the probability of product failure, risk averse consumers have to increase the information search with experience goods. In other words, the impact of the consumer‟s degree of risk aversion on the usage frequency of consumer reviews is greater for experience goods than search goods. Consumers who are more susceptible to informational influence would also increase the consumer review usage when dealing with experience goods (compared to search goods). We expect no moderating effect of product type between the consumer‟s susceptibility to normative influence and the usage frequency of consumer reviews since it is presumable that, for example, the product‟s brand (image), plays a moderating role rather than the distinction between search and experience goods. Based on these findings and logic, we hypothesize:
H5a: The effect of the consumer‟s degree of price consciousness on the usage frequency of consumer reviews is greater for search goods than for experience goods.
H5b: The effect of the consumer‟s degree of risk aversion on the usage frequency of consumer reviews is greater for experience goods than for search goods.
13
2.4 Control Variables
Several demographic variables are included in our model because of their potential influence on the usage frequency of online consumer reviews. Due to the fact that the emphasis of this thesis does not lay on these variables, these demographics are solely included as control variables.
Gender. Early on in the beginning of the internet era, research showed that approximately 95 percent of the internet users were male (Weiser 2000). Although this gender gap in internet use is diminishing, the patterns of internet usage differ from each other. A recent study conducted by Joiner et al. (2012) showed that males significantly make more use of the internet to get information about a product or service than females.
Age. Generally, it is found that increasing consumer age impacts the adoption of online shopping negatively, because younger consumers are more innovative and familiar with computers (Naseri and Elliott 2011). Naseri and Elliott (2011) also confirmed this relationship in their study. A study conducted by Joines, Scherer, and Scheufele (2003) suggested that younger people are more likely to shop online, and research performed by Sorce, Perotti, and Widrick (2005) showed that younger consumers search for more products online than older consumers. We presume that an increase in the use of online shopping results in a higher exposure and use of consumer reviews and, thus, as age increases, the use of online reviews decreases.
14
15 3. Research Methodology
3.1 Data Collection Method
Using an online questionnaire, we will collect our data via various channels (see Appendix A for our questionnaire design). The survey link will be distributed in our immediate environment and through multiple internet discussion forums. Our intention is to get a diverse as possible sample, in the sense of a wide variety of educational levels, wide age range, and respondents from all continents. Tabachnick and Fidell (2007) suggested that the sample size for testing a full regression model should ideally be 8 times the number of independent variables plus 50. For a factor analysis, generally 300 cases are needed. Therefore, our aim is to get at least 300 respondents for our survey. All concepts in our conceptual model are measured altogether with one survey. By doing this, it should be noted that correlations between variables measured with the same method can be inflated because of the action of common method variance (Spector 2006). In regression analyses, common method variance can influence empirical results (Siemsen, Roth, and Oliveira 2010).
3.2 Survey Development, Measurement, and Manipulation
In order to measure the constructs, we base our constructs on existing literature. Some scales were slightly modified to fit our specific situation. Table 1 shows the used items for each construct.
Dependent variable. The usage frequency of online consumer reviews is measured on a 7-point Likert scale with 1 as least frequent and 7 as most frequent use of online reviews. We have chosen a 7-point over a 5-point scale because of the tendency of being too neutral on a 5-point scale (Colman, Norris, and Preston 1997). Using a 7-point Likert scale would prevent responses that are cluttered at the center of the scale.
Price consciousness. Sinha and Batra (1999) developed a measurement scale for category price consciousness. This scale is modified slightly to measure the consumer‟s price consciousness for durable goods as home furniture and cameras (see product type under section 3.3) can be assigned to this category. The authors have tested the scale on internal consistency reliabilities of the latent constructs, followed by a confirmatory factor analysis.
16
Informational and normative influence. Girard (2010) developed scales for measuring both normative and informational social influence. Our scale is based on the items of this study due to the high factor loadings and the performed parsimonious validity and reliability tests.
Product type. Product type was manipulated by randomly assigning respondents to a question involving either a search good or an experience product. Huang, Lurie, and Mitra (2009) noted that the typical search and experience goods classification, originating from the work of Nelson (1970, 1974), may be different in search behavior when comparing the online setting to the traditional retail settings. Therefore, the authors may classify products differently than it traditionally is done. Following the work of Huang, Lurie, and Mitra (2009), cameras were selected as experience goods while home furniture was selected as search good. The authors noted that “consumers shopping for a camera can read extensive product reviews from other consumers and thus can „experience‟ these products before purchase” (p. 56). Product type was included as a dummy variable with 0 as search good and 1 as experience good.
Control variables. Age was measured on a continuous scale, gender on a dichotomous scale, and education on a categorical scale ranging from elementary school to university or higher.
3.3 Validity and Reliability of Measures
To assess the clarity and wording of the questions, the survey was checked by people with different (educational) backgrounds and level of education. In total, 14 people with different educational levels (ranging from secondary vocational to university or higher) and from different fields (medical, technological, and business) were asked to comment on the questionnaire successively. Based on the evaluators‟ remarks, minor changes to the questionnaire were successively made until we did not receive any remarks regarding the clarity and wording of the questions anymore.
17
acceptable, and values greater than 0.50 are generally considered necessary for practical significance (Lederer et al. 2000; Vijayasarathy 2004).
We have conducted the factor analysis in two steps. First, we included all items within one factor analysis to get an overview of the underlying concepts. Second, we have run a factor analysis on each individual concept. The first step revealed that all concepts were derived according to our theoretical constructs, with all communalities being well over 0.50 and with the absence of cross-loadings. In our second set of analyses, all factor loadings were well over 0.50 and the communalities also exceeded the threshold value of 0.50. All factors resulting from the factor analyses are saved on basis of their respective factor scores. The results from the factor analysis are shown in Table 1.
Price consciousness. Price consciousness was measured with four items, explaining 60.76% of the variance. The KMO value was sufficient (0.723) and the Bartlett‟s test of Sphericity was found significant (p < 0.001). The Chronbach‟s alpha was 0.781 and removing an item did not lead to higher internal consistency.
Risk Aversion. Risk aversion is measured with three items (Chronbach‟s alpha = 0.671), explaining 60.58% of the variance. The KMO was sufficient (0.628) and the Bartlett‟s test of Sphericity was found significant (p < 0.001).
Informational influence. Informational influence was measured with three items (Chronbach‟s alpha = 0.760), explaining 67.65% of the variance. Both the KMO and the Bartlett‟s test of Sphericity threshold were met (0.641 and p < 0.001, respectively). Removing an item did increase the Cronbach‟s alpha slightly (0.775). However, due to the theoretical support of the concept, we did not exclude the item from the factor.
18
Concept Items
Commu-nalities Factor loadings Cronbach’s alpha Degree of Price Consciousness
• When it comes to buying a durable good, I rely heavily on price. • When buying a durable good, I look for the cheapest brand available. • Price is the most important factor when I am choosing a durable good.
• When purchasing durable goods, I tend to buy the lowest-priced product that will fit my needs. 0.658 0.611 0.602 0.560 0.811 0.752 0.776 0.748 0.781 Degree of Risk Aversion
• I would rather stick with a brand I usually buy than try something I am not very sure of. •I enjoy taking chances in buying unfamiliar brands just to get some variety in my purchases. •I am very cautious in trying new/different products.
0.705 0.599 0.513 0.840 0.774 0.716 0.671 Susceptibility to Informational Influence
•I would seek information about various brands from an association of professionals or independent group of experts.
•I would seek information from professionals (e.g. pharmacists, doctors, teachers) who work with the product.
•I would seek brand-related knowledge and experience with products (such as how brand A's performance compares to brand B's) from those friends, neighbors, relatives, or work
associates who I think have reliable information about the brands.
0.784 0.680 0.566 0.885 0.824 0.752 0.760 Susceptibility to Normative Influence
•My decision to purchase a particular brand is influenced by the preferences of people with whom I have social interaction.
•To satisfy the expectations of fellow work associates/class mates, my decision to purchase a particular brand is influenced by their preferences.
•The desire to satisfy the expectations, which others have of me, has an impact on my brand choice. 0.795 0.794 0.751 0.892 0.891 0.867 0.859
19
3.4 Model Specification and Method of Analysis
Most of our scales, including our dependent variable, were derived from 7-point Likert scales. Likert scales can be interpreted as metric scales (Likert 1932) which are suitable for linear regression analyses. Therefore, a multivariate linear regression model will be constructed in order to test our hypotheses. The regressed model will look as follows:
Consumer Review Usage = β0 + β1×PC + β2×RA + β3×NORM + β4×INFOR + β5×PT + β6×[PC×PT] + β7×[RA×PT] + β8×[INFOR×PT] + ε,
with βs as the coefficients, PC as price consciousness, RA as risk aversion, NORM as normative influence, INFOR as informational influence, PT as product type, and ε as the residual.
20
21 4. Results
4.1 Data
4.1.1 Sample Description
The majority of our sample was male (90.2% males and 9.8% females). This can be attributed to the fact that the data was collected on forums where predominantly males are active (technology-themed forums and a wristwatch forum). The average age of our sample was 32 (standard deviation: 12.66), ranging from 14 to 73. 62.2% of the participants were from Europe and 31.1% from North America. The vast majority of the respondents were taking or completed an academic degree (i.e. university or higher; 62.2%), followed by higher vocational and high school (19.8% and 11.3%, respectively). No respondents had elementary school as their highest level of education. 94.5% of our sample was either married (or domestic partnership) or single, and their household size averaged 2.86, ranging from 1 to 10. Most participants had no children under 18 in their household (74.4%), followed by 1 and 2 (13.3% and 8.5%, respectively). In our study, we have manipulated product type by measuring the usage frequency of consumer reviews for either a search or experience good. Participants were randomly assigned to one of the product types. In our sample, 53.7% of the participants was assigned to the search good (home furniture) while 46.3% was assigned to the experience good (cameras). As expected, the usage frequency of consumer reviews was higher for the experience good (mean: 5.80; standard deviation: 1.35) than for the search good (mean: 4.32; standard deviation: 1.82). In order to get a sense of the (variation) in the data and variables, we have averaged the original 7-point Likert scale scores with respect to the items according to the results of the factor analysis. The minimum, maximum, mean, and standard deviation of the variables are given in Table 2.
Variable Minimum Maximum Mean Standard Deviation
Consumer Review Usage 1 7 5.010 1.776
Price Consciousness 1 6.25 3.191 1.198
Risk Aversion 1 7 4.158 1.203
Normative Influence 1 7 2.277 1.367
Informational Influence 1 7 4.873 1.237
22
4.1.2 Data Collection
The data was collected via a large-scale, Internet-based survey over a nine-day period of data collecting. The questionnaire has been kept short (completing the questionnaire would approximately take 5 minutes) in order to increase the response. The data was collected through multiple channels of which our immediate environment (a personal Facebook page and at the Zernike Campus, Groningen, the Netherlands) and various Internet forums. The majority of the respondents have been collected through a large Dutch technology-themed forum and an international wristwatch forum. A modest number of respondents were collected at a forum for audio equipment and a forum for products made by Apple Inc.
4.1.3 Data Purification
Prior to the data analysis, the data were checked on irregularities, inconsistencies, and whether or not the data were complete through various rules. In total, 459 respondents had started the survey of which 335 have completed the survey. Out of the 124 cases that have not completed the questionnaire, 81 respondents have solely opened the questionnaire without answering a question or only completed the first block of questions (demographics). The remainder of the cases typically stopped at the second last block of questions where seven questions about normative influence was presented. This may indicate that the respondents were deterred from continuing by the amount of questions. There were no signs that these incomplete questionnaires were due to the survey design, and were therefore removed from the dataset.
23
answers deviated too much from each other (i.e. answered a 1 or 2 on the first question, while the paired question was answered a 6 or 7). Three cases were identified with 3 or more inconsistent answers. Moreover, the IP addresses of the cases did not correspond with their given continent. Due to these highly inconsistencies, these three cases were removed from the dataset. Also, we have checked whether respondents have answered the second last block solely with the same number on the scale, which may indicate that the participant was reluctant in taking the time in answering the last questions. We have identified 24 cases in which the respondents answered those questions with solely a 1. This could either mean that the participants bluntly filled in 1s, or that the participants are not susceptible to normative influence. Due to the fact that both scenarios are plausible, we have not removed these cases. Lastly, we have checked the dataset on duplicates based on the IP addresses. One duplicate was identified of which the answers were inconsistent (aged 25 versus 54 while both had answered that their total household size was 1). These cases were also removed from the dataset. All in all, after purification of the data, 328 valid cases remained, which is sufficiently large when following both criteria posed by Tabachnick and Fidell (2007) for the linear regression model and factor analysis.
4.2 Assumption Tests 4.2.1 Normality Test
24 Figure 2: Residual Plot
accommodate this change, the beta estimates will be reversed (i.e. the beta signs will be reversed) to make further interpretation easier and more intuitive. Table 3 shows the normality test statistics.
Consumer Review Usage
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
Untransformed -0.724**** -0.677**** -0.568**** -0.576**** -0.570**** -0.581**** Transformed (Log + reflected) -0.119**** -0.135*** -0.406* -0.191* -0.201** -0.200** **** p < 0.001, ***p < 0.01, **p < 0.05, *p < 0.1
Table 3: Normality Test Statistics (Skewness Scores and Kolmogorov-Smirnov Test Significance)
4.2.2 Homoscedasticity Test
When visually examining potential heteroscedasticity by plotting the residuals against the predicted values for the dependent variable, we obtained an extraordinary graph where we found seven parallel down sloping lines (see Figure 2). By filtering the dependent variable on different levels of the 7-point Likert scale, we learned that each parallel line indeed resembles one level on the scale. This may be due to the possibility that the variability of the residuals between the values on the scale of our dependent variable is (too) small. In order to statistically test for
25
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
χ2
(5) χ2 (9) χ2 (10) χ2 (11) χ2 (12) χ2 (13)
3.76 5.37 7.87 8.39 9.14 11.93
****
p < 0.001, ***p < 0.01, **p < 0.05, *p < 0.1 Table 4: Breusch-Pagan Test Results
4.2.3 Collinearity Test
In order to assess whether or not the models are dealing with multicollinearity, a collinearity test was added to our regression models. In neither of the models, signs of multicollinearity were present (1.021 < VIF < 2.578). By adding the moderating effects of product type through model 4 to 6, no multicollinearity was added to the models. Table 5 summarizes the VIF value ranges per model.
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
1.052-2.527 1.041-2.563 1.021-2.575 1.021-2.577 1.021-2.578 1.021-2.578
Table 5: Variance Inflation Factor Value Range per Model
4.3 Model Selection
26
product type (model 3), the AIC turned to negative while the BIC decreased with 66%. The AIC can be calculated with the following formula: AIC = -2LLk + 2k, with k as the number of parameters and LLk the log-likelihood value of the k-model. By manually calculating model 3‟s AIC, we indeed obtained an AIC value of -10.247 (AICmodel3 = -2×17.124 + 2×12). The log-likelihood value is a measure of how much unexplained variability there is in the data (Field 2009). Although it is uncommon for a log-likelihood to be positive, it is plausible. For continuous variables, the probability density of a possible value can be positive, which may result in a positive log-likelihood. Based on the AIC and BIC, model 3 once again outperformed all other models, though not by much. Also, the normality test showed that model 3 and 4 were the only models with an acceptable normal distribution of the residuals at a 5% significance level, while model 5 and 6 had an acceptable normal distribution of the residuals at a 1% significance level. Because of the small improvements of model 3 compared to model 6, and the theoretical support of our hypotheses, model 6 was selected when testing the hypotheses. Table 5 summarizes the beta estimates, t-values, significance levels and diagnostics of the models.
4.4 Predictive Validity
To assess the usefulness of the model and whether or not the model is generally applicable, we predictively validated our constructed model using the Root Average Squared Prediction Error (RASPE), also known as Root Mean Squared Error (RMSE), method by splitting up the dataset in two groups (i.e. an estimation and validation sample). The RASPE can be denoted
27
4.5 Hypothesis Testing
We have tested our hypotheses using a multivariate linear regression analysis. Table 7 summarizes our regression results.
H1 is rejected as the effect of the consumer‟s degree of price consciousness is positive while we hypothesized a negative effect. However, the effect did have a marginal significant effect on the review usage (p < 0.1). We find that the more price conscious a consumer gets, the higher the usage frequency of consumer reviews becomes.
Risk aversion was found insignificant (p = 0.554), meaning that the consumer‟s degree of risk aversion does not affect the consumer review usage. Therefore, H2 is rejected.
Although the susceptibility of normative influence had an expected negative impact on the review usage, the effect was found highly insignificant (p = 0.877). Hence, normative influence does not impact the usage frequency of consumer reviews (H3 is rejected).
In line with H4, a significant positive effect was found between the consumer‟s susceptibility to informational influence and the consumer review usage (p < 0.01). Therefore, a high susceptibility to informational influence translates to a higher usage frequency of consumer reviews.
In our study, neither of the main effects were being moderated by product type, since the effects were found highly insignificant (0.472 < p < 0.939). However, we did find a highly significant main effect of product type (p < 0.001). We consecutively added a moderating effect of product type through model 4 to 6 without any significant result. As expected, the AIC and BIC increased due to the addition of these insignificant moderating effects. By adding interaction effects, it is plausible that multicollinearity issues can arise. However, when examining the VIF values, no multicollinearity issues were detected. Therefore, H5a, H5b, H5c, and H5d are rejected.
28
Hypothesis Hypothesized Effect
Independent Variable Hypothesis
Testing Result
H1 Negative Degree of price consciousness Not Supported
H2 Positive Degree of risk aversion Not Supported
H3 Negative Degree of susceptibility to normative influence Not Supported H4 Positive Degree of susceptibility to informational influence Supported H5a Positive Price Consciousness × Product Type Not Supported
H5b Positive Risk Aversion × Product Type Not Supported
H5c Positive Informational × Product Type Not Supported
29
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
B t-value B t-value B t-value B t-value B t-value B t-value Constant 0.474 7.167**** 0.473 7.341**** 0.621 10.337**** 0.621 10.321**** 0.621 10.310**** 0.621 10.281**** Age -0.050 -3.299*** -0.043 -2.797*** -0.048 -3.466*** -0.048 -3.433*** -0.049 -3.472*** -0.049 -3.466** Gender a 0.026 0.512 0.004 0.086 0.028 0.619 0.028 0.627 0.027 0.609 0.028 0.612 Education b Secondary vocational Higher vocational University or higher 0.054 0.035 0.070 0.742 0.634 1.439 0.060 0.063 0.091 0.854 1.158 1.907* 0.084 0.070 0.117 1.321 1.447 2.729*** 0.084 0.071 0.116 1.321 1.447 2.717*** 0.085 0.070 0.117 1.344 1.439 2.725*** 0.085 0.070 0.117 1.344 1.435 2.721*** Price Consciousness 0.046 3.105*** 0.034 2.540** 0.037 2.027** 0.036 1.933* 0.036 1.931* Risk Aversion 0.001 0.081 -0.002 -0.125 -0.002 -0.114 -0.011 -0.590 -0.011 -0.593 Normative Influence -0.004 0.288 -0.001 -0.110 -0.001 -0.093 -0.002 -0.159 -0.002 -0.154 Informational Influence 0.052 3.506*** 0.053 3.960**** 0.053 3.958**** 0.053 3.933**** 0.054 2.933*** Product Type c 0.232 8.895**** 0.233 8.884**** 0.233 8.881**** 0.233 8.866**** Price Consc. × Product Type -0.006 -0.231 -0.004 -0.163 -0.004 -0.162
Risk Aversion × Product Type 0.019 0.718 0.019 0.721
Informational × Product Type -0.002 -0.076
Diagnostics LR ~ χ2 (k) 11.941** 34.213**** 107.293**** 107.349**** 107.886**** 107.892**** R2 0.036 0.099 0.279 0.279 0.280 0.280 Adjusted R2 0.021 0.074 0.256 0.254 0.253 0.251 AIC 75.105 60.833 -10.247 -8.303 -6.840 -4.846 BIC 101.656 102.556 35.269 41.006 46.262 52.049 a
reference category „Female‟
b
reference category „High school‟
c
reference category „Search good‟
****
p < 0.001, ***p < 0.01, **p < 0.05, *p < 0.1
30 5. Discussion
In this study, we examined the effects of shopping orientation, social influence on the usage frequency of consumer reviews, taking the moderating effects of product type into account. The consumer‟s degree of price consciousness had a significant effect on the consumer review usage, while the consumer‟s degree of risk aversion was found to be insignificant. Surprisingly, price consciousness affected the review usage positively, while we hypothesized a negative effect of price consciousness. As mentioned, Seock and Bailey (2008) showed a significant positive relationship between a consumer‟s price consciousness and the degree of online information searches. Although the authors suggested that price conscious consumers are likely to spend time searching for promotional deals and sales, and comparing prices of products offered by different companies, this study shows that these consumers are also more likely to use consumer reviews than consumers who are less price conscious. In the choice process of highly price conscious consumers, not only information about the price is important, but also the consumer reviews as consumer reviews can be a reflection of the best price, given the overall value of the product. Therefore, it is sensible that price conscious consumers make more use of consumer review than those who are less price conscious.
The insignificant outcome of risk aversion on consumer review usage can be attributed to the question design that measures the concept. In our study, we have measured risk aversion broadly by not specifying a product category or brand in order to capture the overall degree of risk aversion. By doing this, participants may have different products in mind in which they may react, in terms of risk averseness, to differently. For example, a participant may not be risk averse when having groceries in mind (i.e. willing to try other products/brands), while in the case of large household appliances, the participant may be a lot more risk averse. This could eventually lead to unwanted variations in the measurement of the concept and eventually resulting in insignificant outcomes.
31
decisions. On the other hand, the consumer‟s susceptibility to normative influence had no impact on the usage frequency of consumer review. This could mean that normative influence is indeed exerted by the consumer‟s immediate environment (i.e. close friends, family or co-workers), which could result in the omittance of consumer reviews altogether.
32 6. Management Implications
Our study creates insights in how the level of online consumer review usage is influenced, which may help managers in the decision of introducing online review systems. Moreover, this provides insights in which type of consumers may value the possibility of reading consumer reviews the most and engage in eWOM the most. Given a positive experience with consumer reviews and a positive product rating, this may in turn, positively influence the purchasing behavior and decisions.
We showed that consumers who are more price conscious will increase their consumer review usage. This also holds for people who are more susceptible to informational influence. If managers have the available information about their customers in terms of customer characteristics, our findings may help managers to decide whether or not introducing a consumer review system is useful or not. For example, if their customers are highly price conscious and highly susceptible to informational influence, having the ability of reading consumers review may be important and useful to them. When consumer reviews are already present, targeting these consumers with (positive) reviews may help increase sales. Also, our findings create interesting insights for companies that are attracting price conscious consumers by offering relatively less expensive products. For these companies, not only price should be important, but also consumer reviews as highly price conscious consumers are more likely to use consumer reviews in greater extent.
The fact that normative influence was unrelated to the usage frequency of consumer reviews may help managers of companies in certain product categories or niches. For example, it is likely that for fashion brands, normative influence will play a significant role among their customers and for these companies, the added value of a review system is rather limited. Managers should, however, be aware that in an online context, consumer review usage is prominent. This is especially the case for experience goods. Retailers or manufacturers that primarily are dealing with experience goods are much more likely coping with consumers who view consumer reviews in a high degree.
33 7. Limitations and Further Research
Several limitations of our study should be acknowledged and could be addressed in future research.
First, our model is a simple representation of reality. A large part of the variance of consumer review usage has not been explained. Adding additional relevant variables to our model can make the model more complete. Burton and Khammash (2010) proposed many motives why consumers read online reviews. Our model can be extended by relating these motives to the extent to which consumers use online reviews. Furthermore, we have used only two shopping orientations while other shopping orientations may influence the review usage. In order to give a better representation of reality, more shopping orientations can be added to our model, such as the consumer‟s importance for service (Kohijoki and Marjanen 2013; Rigopoulou, Tsiotsou, and Kehagias 2008) and product quality (Gehrt et al. 2007).
Second, we saw that risk aversion had no significant impact on the usage frequency of consumer reviews and we believe that this may be due to the question design of the variable. As mentioned before, it may be the case that the participants had different products in mind while answering the questions where they may react, in terms of risk aversion, to differently. For further research, we suggest to specify the concept to one product category or brand, depending on the context.
Third, our model showed extraordinary residual plots consisting of seven parallel lines. We believe this may be because of the small differences between the distances on the 7-point Likert scale when it comes to measuring consumer review usage. We suggest using another scale that is more accurate, such as a (100-point) continuous rating scale. Online surveys eliminate the impracticalities of manually measuring the positions marked on questionnaires (Brace 2004).
Fourth, it should be noted that our sample largely consisted of males. This might have led to biases in our results and by consciously trying to get an even male/female ratio, the sample would get more representative. Males and females have different traits where certain influences on online information search may be different (Joiner et al. 2012; Park, Yoon, and Lee 2009).
34
garden and patio products as search goods. In the case of experience goods, the authors used automotive parts and accessories, and health and beauty products as well.
35 References
Andrews, J. Craig, Richard G. Netemeyer, Scot Burton, D. Paul Moberg, and Ann Christiansen (2004), “Understanding Adolescent Intentions to Smoke: An Examination of Relationships Among Social Influence, Prior Trial Behavior, and Antitobacco Campaign Advertising,” Journal of Marketing, 68 (3), 110-123.
Arnold, Mark J. and Kristy E. Reynolds (2003), “Hedonic shopping motivations,” Journal of Retailing, 79 (2), 77-95.
Bae, Soonyong and Taesik Lee (2011), “Gender differences in consumers‟ perception of online consumer reviews,” Electronic Commerce Research, 11 (2), 201-214.
Bambauer-Sachse, Silke and Sabrina Mangold (2013), “Do consumers still believe what is said in online product reviews? A persuasion knowledge approach,” Journal of Retailing Consumer Service, 20 (4), 373-381.
Batinic, Bernad and Markus Appel (2013), “Mass communication, social influence, and consumer behavior: two field experiments,” Journal of Applied Social Psychology, 43 (7), 1353-1368.
Bearden, William O. and Michael J. Etzel (1982), “Reference Group Influence on Product and Brand Purchase Decisions,” Journal of Consumer Research, 9 (2), 183-194. Beatty, Sharon E. and Scott M. Smith (1987), “External Search Effort: An Investigation
Across Several Product Categories,” Journal of Consumer Research, 14, 83-95. Bettman, James R., Mary Frances Luce, and John W. Payne (1998), “Constructive Consumer
Choice Processes,” Journal of Consumer Research, 25 (3), 187-217.
Bonfield, E.H. (1974), “Attitude, Social Influence, Personal Norm, and Intention Interactions as Related to Brand Purchase Behavior,” Journal of Marketing Research, 74 (4), 379-389.
Brace, Ian (2004), Questionnaire Design. How to Plan, Structure and Write Survey Material for Effective Market Research. Kogan Page: London.
Breazeale, Michael (2009), “Word of mouse,” International Journal of Market Research, 51 (3), 297-318.
Brown, Mark, Nigel Pope, and Kevin Voges (2003), “Buying or browsing? An exploration of shopping orientations and online purchase intention,” European Journal of Marketing, 37 (11/12), 1666-1684.
36
Burroughs, Richard E. and Rajiv Sabherwal (2002), “Determinants of Retail Electronic Purchasing: A Multi-period Investigation,” Information Systems and Operational Research, 40, 35-56.
Burton, Jamie and Marwan Khammash (2010), “Why do people read reviews posted on consumer opinion portals?,” Journal of Marketing Management, 26 (3/4), 230-255. Büttner, Oliver B., Arnd Florack, and Anja S. Göritz (2013), “Shopping Orientation and
Mindsets: How Motivation Influences Consumer Information Processing During Shopping, Psychology and Marketing, 30 (9), 779-793.
Chaudhuri, Arjun (2000), “A Macro Analysis of the Relationship of Product Involvement and Information Search: The Role of Risk,” Journal of Marketing Theory and Practice, 8, 1-14.
Chen, Yubo, Scott Fay, and Qi Wang (2011), “The Role of Marketing in Social Media: How Online Consumer Reviews Evolve,” Journal of Interactive Marketing, 25 (2), 85-94. Chen, Yue-Yang (2012), “Why Do Consumers Go Internet Shopping Again? Understanding
the Antecedents of Repurchase Intention,” Journal of Organizational Computing and Electronic Commerce, 22, 38-63.
Chevalier, Judith A. and Dina Mayzlin (2006), “The Effect of Word of Mouth on Sales: Online Book Reviews,” Journal of Marketing Research, 43 (3), 345-354.
Collins, Katie, Alan Tapp, and Ashley Pressley (2010), “Social marketing and social influences: Using social ecology as a theoretical framework,” Journal of Marketing Management, 26 (13/14), 1181-1200.
Colman, Andrew M., Claire E. Norris, and Carolyn C. Preston (1997), “Comparing Rating Scales of Different Lengths: Equivalence of Scores From 5-Point and 7-Point Scales,” Psychological Reports, 80 (2), 355-362.
Cui, Geng, Hon-Kwong Lui, and Xiaoming Guo (2012), “The Effect of Online Consumer Reviews on New Product Sales,” International Journal of Electronic Commerce, 17, 39-57.
Darden, William R. and Fred D. Reynolds (1971), “Shopping Orientations and Product Usage Rates,” Journal of Marketing Research, 8 (4), 505-508.
37
Dholakia, Uptal M., Richard P. Bagozzi, and Lisa Klein Pearo (2004), “A social influence model of consumer participation in network- and small-group-based virtual communities,” International Journal of Research in Marketing, 21 (3), 241-263.
Dholakia, Utpal M., Suman Basuroy, and Kerry Soltysinski (2002), “Auction or agent (or both)? A study of moderators of the herding bias in digital auctions,” International Journal of Research in Marketing, 19 (2), 115-130.
Field, Andy (2009), Discovering Statistics Using SPSS. London: SAGE Publications Ltd. Floh, Arne, Monika Koller, and Alexander Zauner (2913), “Taking a deeper look at online
reviews: The asymmetric effect of valence intensity on shopping behaviour,” Journal of Marketing Management, 29 (5/6), 646-670.
Gehrt, Kenneth C., Naoto Onzo, Kazuyoshi Fujita, and Mahesh N. Rajan (2007), “The Emergence of Internet Shopping in Japan: Identification of Shopping Orientation defined Segments,” Journal of Marketing Theory and Practice, 15 (2), 167-177. Girard, Tulay (2010), “The Role of Demographics on the Susceptibility to Social Influence:
A Pretest Study,” Journal of Marketing Development and Competiveness, 5, 9-22. Grant, Robert, Rodney J. Clarke, and Elias Kyriazis (2007), “A review of factors affecting
online consumer search behaviour from an information value perspective,” Journal of Marketing Management, 23 (5-6), 519-533.
Hansen, Torben and Jan Møller Jensen (2009), “Shopping orientation and online clothing purchases: the role of gender and purchase situation,” European Journal of Marketing, 43 (9/10), 1154-1170.
Hartman, Katherine B., James B. Hunt, and Carla Y. Childers (2013), “Effects of eWOM valence: examining consumer choice using evaluations of teaching,” Journal of Behavioral Studies in Business, 6, 1-12.
Hennig-Thurau, Thorsten and Gianfranco Walsh (2003), “Electronic Word-of-Mouth: Motives for and Consequences of Reading Customer Articulations on the Internet,” International Journal of Electronic Commerce, 8 (2), 54-74.
Hu, Nan, Indranil Bose, Yunjun Gao, and Ling Liu (2011), “Manipulation in digital word-of mouth: A reality check for book reviews,” Decision Support Systems, 50, 627-635. Huang, Jen-Hung and Yi-Fen Chen (2006), “Herding in Online Product Choice,” Psychology
38
Huang, Peng, Nicholas H. Lurie, and Sabyasachi Mitra (2009), “Searching for Experience on the Web: An Empirical Examination of Consumer Behavior for Search and Experience Goods,” Journal of Marketing, 73 (2), 55-69.
Jensen, Matthew L., Joshua M. Averbeck, Zhu Zhang, and Kevin B. Wright (2013), “Credibility of Anonymous Online Product Reviews: A Language Expectancy Perspective,” Journal of Management Information Systems, 30, 293-323.
Johnson, Eric J., Wendy W. Moe, Peter S. Fader, Steven Bellman, and Gerald L. Lohse (2004), “On the Depth and Dynamics of Online Search Behavior,” Management Science, 50 (3), 299-308.
Joiner, Richard, Jeff Gavin, Mark Brosnan, John Cromby, Helen Gregory, Jane Guiller, Pam Maras, and Amy Moon (2012), “Gender, Internet Experience, Internet Identification, and Internet Anxiety: A Ten-Year Followup,” Cyberpsychology, Behavior, and Social Networking, 15 (7), 370-372.
Joines, Jessica L., Clifford W. Scherer, and Dietram A. Scheufele (2003), “Exploring motivations for consumer Web use and their implications for e-commerce,” Journal of Consumer Marketing, 20 (2), 90-108.
Kim, Jinsoo, Jaejin Lee, and Matt Ragas (2011), “Exploring eWom in Online Consumer Reviews: Experience Versus Search Goods,” Web Journal of Mass Communication Research, 32, 1-1.
Kohijoki, Anna-Maija and Heli Marjanen (2013), “The effect of age on shopping orientation choice orientation types of the ageing shoppers,” Journal of Retailing and Consumer Services, 20 (2), 165-172.
Kuha, Jouni (2004), “AIC and BIC: Comparisons of Assumptions and Performance,” Sociological Methods and Research, 33 (2), 188-229.
Kulviwat, Songpol, Chiquan Guo, and Napatsawan Engchanil (2004), “Determinants of online information search: a critical review and assessment,” Internet Research, 14 (3), 245-253.
Latané, Bibb (1981), “The Psychology of Social Impact,” American Psychologist, 36 (4), 343-356.
39
Lee, Matthew K.O., Na Shi, Christy M.K. Cheung, Kai H. Lim, and Choon Ling Sia (2011), “Consumer‟s decision to shop online: The moderating role of positive informational social influence,” Information and Management, 48 (6), 185-191.
Lee, Min-Young, Youn-Kyung Kim, and Hyun-Joo Lee (2013), “Adventure versus gratification: emotional shopping in online auctions,” European Journal of Marketing, 47 (1/2), 49-70.
Lennon, Sharron J., Minjeong Kim, Kim K.P. Johnson, Laura D. Jolly, Mary Lynn Damhorst, and Cynthia R. Jasper (2007), “A Longitudinal Look at Rural Consumer Adoption of Online Shopping,” Psychology and Marketing, 24 (4), 375-401.
Li, Yong and Jin-Guan Lin (2009), “Heteroscedasticity diagnostics in two-phase linear regression models,” Journal of Statistical Computation and Simulation, 79 (7), 923-938.
Likert, Rensis (1932), “A technique for the measurement of attitudes,” Archives of Psychology, 22 (140), 1-55.
Lui, Raymond R. and Wei Zhang (2010), “Informational influence of online customer feedback: An empirical study,” Database Marketing and Customer Strategy Management, 17 (2), 120-131.
Lord, Kenneth R., Myung-Soo Lee, and Peggy Choong (2001), “Differences in Normative and Informational Social Influence,” Advances in Consumer Research, 28, 280-85. Ludwig, Stephan, Ko de Ruyter, Mike Friedman, Elisabeth C. Brüggen, Martin Wetzels, and
Gerard Pfann (2013), “More Than Words: The Influence of Affective Content and Linguistic Style Matches in Online Reviews on Conversion Rates,” Journal of Marketing, 77, 87-103.
Mägi, Anne W. (2003), “Share of wallet in retailing: the effects of customer satisfaction loyalty cards and shopper characteristics,” Journal of Retailing, 79 (2), 97-106. Maity, Moutusy, Maxwell K. Hsu, and Lou E. Pelton (2012), “Consumers‟ Online
Information Search: Gen Yers‟ Finding Needles in the Internet Haystack,” Journal of Marketing Channels, 19, 49-76.
Malhotra, Naresh K. (2009), Marketing Research: An Applied Orientation. Pearson Education: New Jersey.
40
Mejri, Mohamed, Mohsen Debabi, and Khaled Nasraoui (2012), “Exploratory Analysis of the Shopping Orientation in the Tunisian Context,” International Journal of Marketing Studies, 4 (5), 1-13.
Meuter, Matthew L., Deborah Brown McCabe, and James M. Curran (2013), “Electronic Word-of-Mouth Versus Interpersonal Word-of-Mouth: Are All Forms of Word-of Mouth Equally Influential?,” Services Marketing Quarterly, 34 (3), 240-256.
Mitchell, Vincent-Wayne (1999), “Consumer perceived risk: conceptualization and models,” European Journal of Marketing, 33 (1/2), 163-195.
Naseri, Mohammad Bakher and Greg Elliott (2011), “Role of demographics, social connectedness and prior internet experience in adoption of online shopping: Applications for direct marketing,” Journal of Targeting, Measurement and Analysis for Marketing, 19 (2), 69-84.
Nelson, Phillip (1970), “Information and Consumer Behavior,” Journal of Political Economy, 78 (2), 311-329.
Nelson, Phillip (1974), “Advertising as Information,” Journal of Political Economy, 82 (4), 729-754.
O'Brien, Robert M. (2007), “A Caution Regarding Rules of Thumb for Variance Inflation Factors,” Quality and Quantity, 41 (5), 673-690.
Park, Cheol and Thae Min Lee (2009a), “Antecedents of Online Reviews‟ Usage and Purchase Influence: An Empirical Comparison of U.S. and Korean Consumers,” Journal of Interactive Marketing, 23 (4), 332-340.
Park, Cheol and Thae Min Lee (2009b), “Information direction, website reputation and eWOM effect: A moderating role of product type,” Journal of Business Research, 62, 61-67.
Park, Jooyoung, Yeosun Yoon, and Byungtae (2009), “The Effect of Gender and Product Categories on Consumer Online Information Search,” Advances in Consumer Research, 36, 362-366.
Paul, Sudhir R. and Xuemao Zhang (2010), “Testing for normality in linear regression models,” Journal of Statistical Computation and Simulation, 80 (10), 1101-1113. Podsakoff, Philip M., Scott B. MacKenzie, and Jeong-Yeon Lee (2003), “Common Method
41
Purnawirawan, Nathalia, Patrick De Pelsmacker, and Nathalie Dens (2012), “Balance and Sequence in Online Reviews: How Perceived Usefulness Affects Attitudes and Intentions,” Journal of Interactive Marketing, 26 (4), 244-255.
Racherla, Pradeep and Wesley Friske (2012), “Perceived „usefulness‟ of online consumer reviews: An exploratory investigation across three service categories,” Electronic Commerce Research and Applications, 11, 548-559.
Rigopoulou, Irini D., Rodoula H. Tsiotsou, and John D. Kehagias (2008), “Shopping orientation-defined segments based on store-choice criteria and satisfaction: an empirical investigation,” Journal of Marketing Management, 24 (9/10), 979-995. Robson, Karen, Mana Farshid, John Bredican, and Stephen Humphrey (2013), “Making sense
of online consumer reviews: a methodology,” International Journal of Market Research, 55 (4), 2-13.
Rose, Susan and Phillip Samouel (2009), “Internal psychological versus external market driven determinants of the amount of consumer information search amongst online shoppers,” Journal of Marketing Management, 25 (1/2), 171-190.
Seock, Yoo-Kyoung and Lauren R. Bailey (2008), “The influence of college students‟ shopping orientations and gender differences on online information searches and purchase behaviours,” International Journal of Consumer Studies, 32 (2), 113-121. Siemsen, Enno, Aleda Roth, and Pedro Oliveira (2010), “Common Method Bias in
Regression Models With Linear, Quadratic, and Interaction Effects,” Organizational Research Methods, 13 (3), 456-476.
Simpson, Penny M., Judy A. Siguaw, and John W. Cadogan (2008), “Understanding the consumer propensity to observe,” European Journal of Marketing, 42 (1/2), 193-221. Sinha, Indrajit and Rajeev Batra (1999), “The effect of consumer price consciousness on
private label purchase,” International Journal of Research in Marketing, 16 (3), 237-251.
Smith, Gerald E. (2000), “Search at different price levels: the impact of knowledge and search cost,” Journal of Product and Brand Management, 9 (3), 164-178.
Sorce, Patricia, Victor Perotti, and Stanley Widrick (2005), “Attitude and age differences in online buying,” International Journal of Retail and Distribution Management, 33 (2), 122-132.
42
Sridhar, Shrihari and Raji Srinivasan (2012), “Social Influence Effects in Online Product Ratings,” Journal of Marketing, 76 (5), 70-88.
Stone, Gregory P. (1954), “City Shoppers and Urban Identification: Observations on the Social Psychology of City Life,” American Journal of Sociology, 60, 36-45.
Tabachnick, Barbara G. and Linda S. Fidell (2007), Using Multivariate Statistics. Boston: Allyn and Bacon.
Taylor, James W. (1974), “The Role of Risk in Consumer Behavior,” Journal of Marketing, 38 (2), 54-60.
Tseng, Shun-Yao (2013), “Risk-taking in Investment Choices: Testing the Effects of Information Search,” Australian Journal of Basic and Applied Sciences, 7 (4), 580-585.
Vijayasarathy, Leo R. (2004), “Predicting consumer intentions to use on-line shopping: the case for an augmented technology acceptance model,” Information and Management, 41 (6), 747-762.
Vincent, Michael (2013), “Fishy Statistics: In Re Consolidated Salmonid Cases and the Problem of Autocorrelation,” Journal of Law, Science and Technology, 53 (2), 143-162.
Weiser, Eric B. (2000), “Gender Differences in Internet Use Patterns and Internet Application Preferences: A Two-Sample Comparison,” CyberPsychology and Behavior, 3 (2), 167-178.
Wodzicki, Katrin, Eva Schwämmlein, Ulrike Cress, and Joachim Kimmerle (2011), “Does the Type of Anonymity Matter? The Impact of Visualization on Information Sharing in Online Groups,” Cyberpsychology, Behavior, and Social Networking, 14 (3), 157-160.
Wooten, David B. and Americus Reed II, “Informational Influence and the Ambiguity of Product Experience: Order Effects on the Weighting of Evidence,” Journal of Consumer Psychology, 7, 79-99.
Wu, Guohua (2011), “Country image, informational influence, collectivism/individualism, and brand loyalty: exploring the automobile purchase patterns of Chinese Americans,” Journal of Consumer Marketing, 28 (3), 169-177.
43
