Faculty of Economics and Business
MSc Marketing
Master Thesis
Will content of product sharing blog
influence share intention?
2
Table of Contents
1. Introduction ... 1
2. Literature review and hypothesis ... 3
2.1
Product sharing blogs as an effective marketing tool through eWOM ... 3
2.2
Content of product sharing blogs ... 4
2.2.1
Novelty of content... 4
2.2.2
Usefulness of content ... 5
2.2.3
Attractiveness of content... 5
2.3
Popularity of bloggers ... 6
2.4
Motivations to share ... 6
3. Hypotheses... 7
3.1
Novelty of content and share intention ... 7
3.2
Usefulness of content and share intention ... 8
3.3
Attractiveness of content and share intention ... 9
3.4
Moderation effect: popularity of blogger as a moderator ... 10
4. Conceptual model ... 12
5. Methodology ... 12
5.1
Data collection... 12
5.2
Questionnaire design ... 13
5.3
Manipulation of blogs ... 14
6. Results ... 17
6.1
Reliability test of the product (digital cameras) ... 17
6.2
Demographics descriptive statistics ... 17
6.3
Validity test of items ... 19
6.4
Reliability test ... 21
6.5
Variables‘ descriptive statistics and correlations ... 22
6.6
Hypothesis testing ... 23
6.6.1
Manipulation check of conditions ... 23
6.6.2
Regression analysis ... 24
7. Discussion and implications ... 26
7.1
Summary of results... 26
7.2
Implications ... 28
3
7.2.2
Implications for practice ... 29
7.3
Limitations ... 31
8. References ... 32
Appendix 1 Questionnaire ... 41
Appendix 2 Blog sample:... 43
1 Abstract
Writing product sharing blogs is increasingly popular. Among these blogs, some sponsored product sharing blogs have strong electronic word-of-mouth (eWOM) effect through sharing, and content of the blogs plays a crucial role in affecting the sharing behavior. In the article, an empirical 2 (novel vs. less novel)* 2 (useful vs. less useful)* 2 (attractive vs. less attractive) between design experiment was conducted to explore the effect of the three content relevant attributes of product sharing blogs on individuals‘ share intention. Also, the moderation effect by the popularity of bloggers was examined. Results of this empirical study showed that novelty, usefulness and attractiveness of content are positively related to individuals‘ share intention. However, the moderator popularity of blogger had no effect on previous relationships. Implications and limitations of the research were also discussed.
Key words: content relevance, product sharing, blogs, popularity of blogger, share intention
1. Introduction
In recent years, it is increasingly popular for people writing blogs. According to social media study by Universal-McCann (2012), the percentage of globally active internet users writing blogs has increased from 28 percent in 2006 to 50 percent in 2011. Among them, people are fond of expressing their attitudes of products in their blogs. The Blogosphere 2011 report by Technorati shows that 38 percent of bloggers (author of blogs) share a product or a brand in their blogs1. This brand-related content is important message searched by consumers because it helps them making purchase decisions (Chu & Kim 2011). Some companies noticed the phenomenon and began to use the electronic word-of mouth (eWOM) content to promote products (Chevalier & Mayzlin 2006; Goldsmith & Horowitz 2006). eWOM refers to ―any statements made by potential, actual, or former customers about a product or company, which is made available to a large amount of people and institutions via the Internet. (Hennig-Thurau et al., 2004) Blog is one form of eWOM (Osman, Yearwood and Vamplew 2009; Litvin, Goldsmith and Pan 2008; Chen, Shang and Li 2014). Thus, product sharing blog can function as a marketing tool.
2
Normally, companies provide free product samples to some popular bloggers and then these bloggers record, evaluate and share product usage feelings in their blogs to influence their readers. For example, Michelle Phan, a popular professional makeup artist sponsored by Lancôme (a beauty product company under L‘Oreal Group), has over 5 million readers of her blog and the page views of her newest product post published in 11 February 2014 after only one night are 1430292. Clearly, eWOM can efficiently contact far-reaching people (Hung &
Li 2007; Gruen, Osmonbekov and Czplewski 2006).
Most prior studies just focus on utilizing blogs to increase consumers‘ brand knowledge (Sinha & Ahuja 2011) or as a less important way to assist gain competitive advantages for companies (Huang, Yung and Yang 2011) or a low-cost way to assess their service quality (Pan, MacLaurin and Crotts 2007). However, less is paid attention to the content relevant study of blogs implementing eWOM.
Content of blogs plays a significant role in attracting a large amount of people reading, especially in today‘s knowledge-driven economic society (Grant 1996). The well known slogan ‗Content is king‘ is also accepted by virtual community (Huizingh 2000; Jacobs, 2000) and it is the most important characteristic defining the virtual community (Gray 1999). However, the value of content itself is limited if it was not transmitted among people in virtual communities and thus the effectiveness of eWOM will be influenced. Specifically, the effect of eWOM is determined by the supply of content. Through a large amount of content supply (content sharing), virtual community can be fostered and thus users can satisfy their expectations and needs, such as enriching knowledge, seeking support and making friends (Andrews 2002; Zhang & Hiltz 2003). Also, companies will profit from eWOM to increase the interactivity with users (Deighton and Kornfeld 2009) or implement more effective marketing plans (Marchiori, Milwood and Zach 2013).
Manifestly, the key to content supply rests on whether users would like to share knowledge with other members (Chiu, Hsu and Wang 2006). Thus, exploring readers‘ share intention toward content is meaningful and essential. However, some researchers just focus on the effect of emotion aspect of content transmission (Berger and Milkman 2012) other than
3
content relevant attributes on share intention. What‘s more, there are no relevant studies specific in products sharing aspect.
The content relevant attributes refers to subjective information relevance (Xu & Chen 2006). According to information rules by Shapiro, Varian and Becker (1999), usefulness, attractiveness and novelty are key elements of information relevance and they determine whether the information is salient for people in such an information explosion era and whether it can boost information economy.
Therefore, the article explored the three attributes of content of product sharing blog which may affect readers‘ share intention and examine the assumed moderation effect of popularity of bloggers on previous relationships respectively because readers prefer to seek for reputable sources when they make decisions (Lutz 1985; Mackenzie and Lutz 1989). Then, a conceptual model was proposed in the following. The study mainly adopted online questionnaires to collect data and tested the influential elements of content that affect individuals‘ share intention. The results provide implications not only for virtual communities increasing website traffic but also for product managers better implementing effective blog marketing and for marketing analysts seeking out the optimal webpage of companies. In addition, it also gives some insights for non-sponsored product sharing bloggers enhancing their popularity and gaining more profits.
2. Literature review and hypothesis
2.1 Product sharing blogs as an effective marketing tool through eWOM
4
to promote their products and services (Chiang & Hsieh 2014; Carson 2008; Lin & Huang 2006; Huang 2012), build public relations (Yang & Lim 2009), increase consumer brand knowledge (Sinha & Ahuja 2011), obtain a competitive advantage (Huang, Yung and Yang, 2011) and can also be a cost-effective way to assess service quality and improve overall performance (Pan, MacLaurin and Crotts 2007).
Product sharing blog is a kind of third-party sponsored blog. The authors (bloggers) evaluate and share product usage feelings or visual effects in order to recommend and promote the products to their readers. In the blogs, except for text, other multi-media are also used, such as photos, videos or audio. Through these media, bloggers can better display products‘ functions and effects to readers. After reading, readers can leave comments or discuss with each other below and can even share the blog to their friends or strangers if they would like to, which makes the blog as a virtual discussion community. In this respect, blog can be seen as a social media that conveys user generated content and thus is a type of eWOM (Osman, Yearwood and Vamplew 2009; Litvin, Goldsmith and Pan 2008; Chen, Shang and Li 2014; Xiang & Gretzel 2010).
2.2 Content of product sharing blogs
―Content is king‖ is a well-known slogan and it is accepted by virtual community (Huizingh 2000; Jacobs, 2000). Gray (1999) even defines content as the most crucial characteristic defining virtual community. Blog as a type of virtual community can execute eWOM and its content plays a significant role in attracting a large amount of people reading, especially in today‘s knowledge-driven economic society (Grant 1996). Hence, the attributes of blog‘s content deserve attention. The illustrations of content relevant attributes novelty, usefulness and attractiveness will be elaborated respectively in the following.
2.2.1 Novelty of content
5
show different behaviors (e.g. knowledge enriching). Obviously, individuals may have different perceptions of novelty towards different conditions. Berlyne (1970) and Chen, Shang and Li (2014) have also advocated that the more familiar with a stimulus, the less novel people would feel. Specifically, one may feel less novel if she/he had seen the content of the blog before, so novelty is compared to past experience (Pearson 1970). Moreover, novelty can also be compared between individuals because different people may show different feelings of novelty when facing the same new content (Harter 1992). Thus, Novelty is an important content relevant attribute in this model to explore people‘s share intention and it can be defined as ―the extent to which the content of blogs is perceived new by the reader‖. 2.2.2 Usefulness of content
Usefulness can be another important factor that will influence individuals‘ attitude and the subsequent behavior. (Davis 1989) Specifically, individuals would like to share information only if the content is useful (Lu & Hsiao 2007). Clearly, it is also a cognitive attribute. Davis (1989) defined the concept ―perceived usefulness‖ as ―the degree to which a person believes that utilizing a particular system would enhance his or her performance‖ in his technology acceptance model (TAM). Product sharing blog presents the instructions and functions of products to readers, and they can acquire and learn some information about products. For example, people can know what functions the product have, how to use the product or which product is suitable for them, and finally they can perceive whether the information is useful or not. The attribute will also be used in this article and the concept of usefulness can be defined as ―the extent to which a reader believes that reading the content of product sharing blogs will enhance their knowledge about the products‖.
2.2.3 Attractiveness of content
6
also be applied to blog context. When readers view the content of blogs, they will have their aesthetic perceptions and then form their attitudes and intentions (Bandura 1986, 1989). In detail, the attractiveness of content refers to aesthetic aspect which could stimulate individuals‘ interest or desire to make decisions.
Therefore, attractiveness is also an important content relevant attribute included in this model and the concept of attractiveness can be defined as ―the extent to which a reader believes that the content in product sharing blogs would enhance his/her aesthetic sense and arouse his/her sense of interest.
2.3 Popularity of bloggers
Advertisers or marketers normally use popular endorsers (endorser effect) as their spokespersons to support their advocacy, and the effect is also frequently used in product extension. These cases have been verified by many researchers. Deutsch and Ross (2003) concluded that reputable directors are signals for entrepreneurial firms‘ (young firms) success. New products linked to the existed successful reputable products on the market will be more likely accepted by consumers. (Choi & Scarpa 1992; Herbig & Milewicz 1995) The reason is that individuals believe information from popular subjects are more credible, and people prefer to seek for credible sources when they make decisions (Lutz 1985; Mackenzie & Lutz 1989)，so they are more likely to rely on popular subjects‘ opinions. In addition, highly credible sources will be more trusted by people (Jain & Posavac, 1999; Park & Lee 2009) and their trust will further influence their attitudes and intentions (Suh and Han 2003; Wu and Chen 2005; Lee, Lee, Park and Han 2011). Product sharing bloggers generally are popular bloggers sponsored by companies. They have a large amount of followers (fans) and also have professional knowledge or skills to utilize and display products, so the degree of popularity will become a signal of whether the blogger can be trusted by readers and may further influence their share intention.
Therefore, the variable popularity of bloggers, as an external factor of the model, will be regarded as a moderator in this article and will be examined whether or not enhance the relationship between content relevant attributes of blogs and share intention.
2.4 Motivations to share
7
virtual communities and thus the effectiveness of eWOM will be influenced, because the supply of knowledge (e.g. content) determines whether it fosters the virtual community, and the key to the supply is the willingness to share knowledge with other members (Chiu, Hsu and Wang 2006).
However, willingness to share depends on motivation (Ajzen, 1985, 1991). In other words, individuals will not voluntarily engage in knowledge sharing under a no reward virtual community. Lacking of motivation from a knowledge contributor impedes the knowledge sharing (Chiu, Hsu and Wang 2006) Hence, share motivation should be foremost considered. According to the Social Cognitive theory (Bandura 1986; 1989), individuals prefer to engage in the behaviors that are perceived to result in favorable consequences. These favorable consequences include both community-related and personal-related in terms of knowledge sharing (Chiu, Hsu and Wang 2006), because knowledge contributors have outcome expectations toward the sharing behavior (Compeau & Higgins 1995). Specifically, contributors expect to acquire positive results of the community or themselves through sharing behavior. Several studies have suggested that individuals share knowledge with the expectation of helping the virtual community to accumulate its knowledge, continue its operation, and grow (Bock & Kim 2002; Kolekofski & Heminger 2003; Lesser 2000). In addition, Andrews (2002) and Zhang & Hiltz (2003) demonstrated that individuals would like to share knowledge within virtual communities with the expectations of enriching knowledge, seeking support and making friends. Butler, Kiesler and Kraut (2013) also suggested that the main reason for individual to share knowledge is their expectation of being seen as skilled, knowledgeable or respected. Clearly, only content satisfying readers‘ outcome expectations can blogs be shared by people.
3. Hypotheses
3.1 Novelty of content and share intention
8
(Flavell 1977) and this behavior has also been regarded as an indispensible motivator pleasuring readers in many cases (Chen, Shang and Li 2014；Assaker, Vinzi and O‘Connor 2011) Novelty , on one hand, can be new and potentially discrepant information (Rogers 1962; Fiske & Maddi 1961; McClelland 1955). on the other hand, it can be object variations (Hirschman 1980), which means that individuals prefer to switch their use among objects that are already known before because of boredom or fatigue other than to obtain additional new information. If the content of product sharing blog are perceived familiar or perceived no variation, readers may find it bored. Conversely, if they perceive the information to be new, they will be more enjoyable to read the blog. Chen, Shang and Li (2014) and Carlile (2004) also indicated that one reason for people sharing knowledge is seeking novelty. Fredrickson (1998, 2001, 2002, and 2003) further affirmed that people will have positive emotion when they see novel objects, and the more positive the emotion, the more likely for people to share (Berger & Milkman 2012).
Normally, there are a collection of new arrivals produced during a fixed period and these products would be recommended simultaneously in the blog, so the content would be new to readers. Also a new way for displaying these products would be recommended by these sponsored popular bloggers through multi-media such as photos and videos. Readers may perceive the content itself or the way of product displayed as novel, then they may have more positive emotions and therefore tend to share the blog. Thus, the hypothesis is proposed below:
H1: The more novel perceived by readers of the content in product sharing blog, the more likely for them to share.
3.2 Usefulness of content and share intention
9
Riedl 1998) because it has social exchange value (Homans 1958).
Several researchers have affirmed that share intention is closely related to these individual needs towards useful information. Wojnicki & Godes (2008) empirically indicated that individuals who prefer to appear knowledgeable are more likely to share useful information. In addition, Lee and Lee (2010) also proposed that people are more likely to share when the content is useful for others.
Product sharing blog presents basic product information and visual effects to readers. Obviously, the detailed and concrete information described in the blogs will tell people how to use the products and some relevant usage feelings or effects and then they will know more about the products. After reading the content, people can perceive whether the usefulness of information could enhance their knowledge or whether it is valuable for others. Thus, the assumption is that readers who perceive the content as more useful are more likely to share the blog. The hypothesis is proposed below:
H2: The more useful perceived by readers of the content in product sharing blog, the more likely for them to share.
3.3 Attractiveness of content and share intention
Attractiveness refers to a quality that causes an interest or desire in something or someone. (Ortony 1990) In reality, between two identical objects, the more attractive one is always preferred and is perceived to work better. Some researchers have verified the fact. Kurosu and Kaori (1995) designed a field experiment with two forms of automated teller machines (ATM) with totally identical functions (same buttons and layout), but one had the buttons and screens arranged attractively and the other arranged unattractively. The result shows that people (Japanese) think the attractive one is easier to use. The same experiment was also examined by Tractinsky (1997) in Israel because he thinks cultural differences may result in biased results due to culturally dependent aesthetic preferences. However, the results replicated with the Japanese one and even stronger in Israel than that in Japan.
10
cognitive responses and help decision making (Isen et al., 1985; Isen & Daubman 1984; McDuff 2012). Clearly, emotion functions as a link between attractiveness perception and intention.
Santayana (1955), Maritain (1966) and Tatarkiewicz (1970) pointed out that attractive objects can directly and immediately convey pleasure sense (positive emotion) without any reasoning, and aesthetic pleasure is an important need people pursue (Vogt, Fesenmaier and MacKay 1994) and it is also the motivation and gratification for people reading (Nell 1988). Moreover, several studies have also indicated that emotion plays an important role in influencing people‘s share intention. Heath, Bell and Stemberg (2001) indicated that emotional aspects of content affect whether information is shared. In addition, people may share emotionally charged content to satisfy their needs and expectations, such as expressing feelings (Rime et al. 1991), reducing dissonance (Festinger 1957) and consolidating social-network building (Peters & Kashima, 2007). Among them, the content that evoking high-arousal positive emotions will be more likely shared. (Berger & Milkman, 2012)
In a product sharing blog, artistic displays of products and aesthetic visual effects will arouse readers‘ interest and enhance their aesthetic sense and then they are more likely to have positive emotion, finally they would like to share the blog with others. Thus, we assume that attractive content is positively related to share intention and the hypothesis is proposed as followed:
H3: The more attractive perceived by readers of the content in product sharing blog, the more likely for them to share.
3.4 Moderation effect: popularity of blogger as a moderator
11
bloggers could bring much more benefits or positive outcomes to them than those who are less trusted (less popular). Obviously, the concept is also consistent with that individuals would like to engage in those behaviors that they expect to produce positive outcomes (Bandura 1986; 1989).
The three attributes (novelty, usefulness and attractiveness) of content will be perceived by individuals based on their expectations of the positive outcomes. The empirical research conducted by Chrysochou and Krystallis (2013) has concluded that the ad messages of fashion goods are perceived as more novel, useful and attractive when the ads are from a reputable source (company-created) versus a less reputable source (consumer-generated) because consumers believe the ad message could give them professional advices. The study by Agrawal and Kamakura (1995) has also affirmed that investing in celebrity endorsement in advertising is worthwhile (positive stock returns) because consumers perceive the advertisement as more useful to acquire brand knowledge. In addition, Kamins (1990) indicated that the product (car) exposed to consumers would be perceived as more attractive when utilizing a reputable endorser, and the effectiveness of advertising would be enhanced. Thus, it can be assumed that when the blogger is more popular, readers‘ perception toward the three attributes of the content of blogs would be much more positive due to positive outcome expectations and thus, they are more likely to share the blog. The hypotheses are proposed respectively below:
H4a: Popularity of bloggers will enhance the relationship between novelty of content of the product sharing blog and share intention.
H4b: Popularity of bloggers will enhance the relationship between usefulness of content of the product sharing blog and share intention.
12
4. Conceptual model
Figure 1 illustrates the research model, in which it is hypothesized that a reader‘s willingness to share blog is determined by the content of product sharing blog in three aspects: novelty, usefulness and attractiveness. Additionally, popularity of blogger was proposed as a moderator in the research model.
5. Methodology
5.1 Data collection
Virtual communities have flourished online and people frequently participate in discussions of diverse issues (Horrigan 2001; Haythornthwaite 2001), so online survey is an appropriate way to collect data, because it is cost-effective and time-saving and, most importantly, it can reach specific population with specific interest (Wright 2005). In order to collect the data of the readers who are interested in reading product sharing blogs, the link of the survey outlining the goal of the study was posted on two well known product sharing blog websites Sina weblog/blog 3 and Zol blog4. Sina is one of the mainstream and popular blog information
platforms of China.5 Its monthly active users had reached 129.1 million till December 2013
13
and the total number of blogs posted has exceeded 280 million in December, 20136. Also, Zol
is another one of biggest product sharing blog spheres in China and it has 11 million registered users and 950000 active users sharing product reviews in the forum7.
In addition, a QQ (chat software) friend list was also used to ask friends who are interested in reading product sharing blogs. Only those responding ―yes‖ were invited to fill in the online questionnaire. Also, the students in the university were invited to fill in a paper questionnaire if they are interested in reading product sharing blog as well.
In order to enhance respondents‘ participation and the accuracy of the answer, a thirty euro Amazon voucher was made to attract them and stated that ―if they completed the questionnaire they may have a chance to win the voucher‖. Before a pretest, 8 subjects were asked to check the expression of items (content validity), because the wording of questions should at least have face validity (Nunnally & Bernstein, 1994, chap. 3).
Then, 30 respondents who are interested in reading product sharing blog were invited to fill in the online questionnaire and some modification advices and feedbacks were obtained from them and the final version of the online questionnaire was formed (Appendix 1).
5.2 Questionnaire design
In the questionnaire, there are two parts obtaining information. The first part is the demographic profile, which asked respondents basic profile, such as gender, age, income and the time period of reading product sharing blogs. The second part is the main construct including the questions of three content relevance attributes. According to previous studies (Xu & Chen 2006; Davis 1989; Rossiter 2002; Kurosu & Kashimura 1995; Tractinsky 1997), a five-likert scale ranging from (1) ―strongly disagree‖ to (5) ―strongly agree‖ was used to measure the attributes and questions were updated to the research setting (digital cameras) because it was used as the product recommended in the product sharing blogs (see Appendix 1). The specific questions of corresponding attributes are explained below.
As for novelty, Xu and Chen (2006) designed a generic measurement scale for the attribute, and it can be used here. In their research, they proposed that items for measuring novelty can from four aspects of content: 1) amount of new information, 2) the degree of knowing the
14
phenomenon or arguments, 3) amount of unique information come across for the first time and 4) its similarity to previous content.
As for usefulness, Davis (1989) conducted six items of the attribute through using a specific system. In addition, Hsu et al. (2013) developed three items of the attribute toward online shopping intention, and they are 1) performance improvement, 2) effectiveness increase and 3) productivity enhancement, overlapping three items of previous six-item design. Therefore, the items of content usefulness were adapted based on Davis (1989).
As for attractiveness, both Kurosu & Kashimura (1995) and Tractinsky (1997) designed only one item of attractiveness in aesthetic aspect, namely how beautiful an object is. However, more items of the attribute were formulated in this article because except for text, multi media such as photos, audio and video are also used in the blog and these elements will boost vividness of product presentations, which can attract higher levels of attention from consumers than less vivid ones (Kisielius and Sternthal 1984, 1986), so the items were designed to ask respondents‘ perception of the images. Finally, a five-item scale was designed to measure the attribute.
Popularity of bloggers as an external variable is an influential factor for readers choosing blogs, because some characteristics of bloggers, such as professional skills, vocation or physical attractiveness, may influence readers‘ perceived credibility. Thus the variable is included as a moderator in the research model. In the questionnaire, a short introduction of the blogger (popular or less popular) was provided and the context also tells that the blog provided to respondents was written by the blogger.
Willingness to share is defined as ―the intention to share the product sharing blog‖ in the article. The dependent variable will be designed in five-point scale as well. In addition, a question asking readers whether they are interested in digital cameras was included to make sure all the respondents are valid (deleting those who are not interested in digital cameras). Also an open question was added to ask readers why they like to read product sharing blog, which could help further explore readers‘ share motivations.
5.3 Manipulation of blogs
15
example, novelty has novel and less novel two levels, same with usefulness and attractiveness (See figure 2). Therefore, a 2 (novelty: novel vs. less novel) x 2 (usefulness: useful vs. less useful) x 2 (attractiveness: attractive vs. less attractive) factorial design was conducted, resulting in 8 experimental conditions (8 blogs). Also, the moderator, popularity of the blogger, is also divided into two levels: popular vs. less popular, which finally results in 16 conditions of the model in total (Table 1). In addition, between-design was adopted so every participant was served in only one condition, which avoids the possibility of previous conditions giving hints to the latter conditions (Aronson, Wilson and Brewer 1998).
Table 1.Sixteen combinations of blog
Groups Combinations of blog Popular
group
1. Novel & useful & attractive 2. Novel and less useful & attractive 3. Less novel & useful & attractive 4. Less novel & less useful & attractive 5. Novel & useful & less attractive 6. Novel & less useful & less attractive 7.Less novel & useful & less attractive 8.Less novel & less useful& less attractive Less
popular group
9. Novel & useful & attractive 10. Novel and less useful & attractive 11. Less novel & useful & attractive 12. Less novel & less useful & attractive 13. Novel & useful & less attractive 14. Novel & less useful & less attractive 15.Less novel & useful & less
attractive
16
In order to test three attributes of product sharing blog on share intention, a specific product was required. A digital camera was finally chosen as the product recommended in the eight blogs, because it is a neutral product and can avoid gender bias (Hong and Toner 1989). The topic of the blog is about using the digital camera to take good photos, so all the manipulations of attributes of blog actually are specific manipulations of digital cameras (research setting). Specifically, in order to test the novelty of content, a latest Canon digital camera (EOS 70D) with new function descriptions of EOS 70D was recommended in the more novel condition and an old edition Canon digital camera (EOS 550D) without function descriptions was introduced in the less novel condition. As for usefulness attribute, thorough vs. simple descriptions and procedures were differentiated in more useful and less useful condition correspondingly. Also, high-definition (more attractive) vs. low-definition (less attractive) images were provided to test the influence of attractiveness attribute. The specific designs of the sixteen blogs are presented in Appendix 2 and the interpretations of each level of the attributes can be seen in Table 2.
Table 2． Levels of attributes interpretation in the blogs Independent
variables
levels Items of levels included in blogs
Novelty Novel A latest Canon digital camera – EOS 70D New function description
Less novel An old edition Canon digital camera – EOS 550D Usefulness Useful Thorough and specific procedures to use the camera
Thorough product description
Useful skills of how to take good photos Less useful Simple and vague procedures to use the camera
Simple product description
Attractiveness Attractive High definition and high quality images Less
attractive
Low definition and low quality images
17
The survey lasted for 2 weeks. The questionnaires of each blog were filled by at least twenty-five respondents, and finally four hundred twenty-four respondents were gained during this period (357 from online forum, 39 from QQ friend list and 28 from the university). Of the four hundred twenty-four samples, four hundred participants are valid. The criterion of deleting depends on the multiple completions of questionnaires (8 respondents), choosing more than one option (3 respondents), duplication of IP addresses (2 respondents) and missing data (11 respondents).
6. Results
6.1 Reliability test of the product (digital cameras)
Before analyzing data, it is required to test the reliability of the product (digital camera) recommended in the blog to avoid gender bias, because whether the answers of male and female toward the digital camera are different will influence further analysis. Thus, One-way Analysis of Variance (One-way ANOVA) of the questions on the entire 400 samples (two groups: male and female) was used to test whether means of different groups (genders) are equal (null hypothesis) through analyzing the corresponding probability of F-value under 95% significance level (Fisher 1922). Specifically, if the probability (p-value) < 0.05, null hypothesis is rejected, indicating that means for male and female answering these questions are different. However, the result in Table 3 (F-value = 1.32, P-value = 0.18 > 0.05) shows that there is no differences of the scores of these questions between male and female, implying that both male and female hold the same attitude toward the digital camera. Thus, adopting a digital camera (neutral product) in the experiment setting is a right choice and further analysis can be conducted.
Table 3：One Way ANOVA of 400 samples
Sum of Square df Mean Square F P-value
Between Groups 32.72 14 2.34 1.32 0.18
Within Groups 10565.84 5985 1.77
Total 10598.56 5999
6.2 Demographics descriptive statistics
18
the survey. Among these people, their ages mostly are from 26 to 45, and age group between 26 and 35 accounts for the highest percentage (42.5%) and 36-45 (29%) ranks the second highest. In education item, the majority of respondents are bachelor (204) and master (132), occupying 51% and 33% respectively, implying that most people are well educated. In income item, the sample mainly focuses from 3000 to 7000 Yuan, accounts for 70.4% in total. Among them, 3000-3999 (65), 4000-4999 (100), 5000-5999 (65), 6000- 6999 (51) constitutes 16.3%, 25%, 16.3% and 12.8% respectively, showing that 70.4% of respondents have high level income. Seen from the length of blog reading in years, 276 respondents (69%) have been reading blogs for more than 2 years. Among them, most respondents read blogs between 2 and 3 years (142, 35.5%). Above information shows that most respondents are well educated and have high income, and are fond of reading blogs and these figures are in line with the Blogosphere 2011 report by Technorati that people who are fond of reading blogs (80% have been reading blogs for two years) are affluent and well educated (79% have bachelor degrees). Therefore, the data of demographics are sufficiently reliable.
Table 4 Demographics descriptive statistics
Questions Items Frequency Percent
19
￥5000-5999 65 16.3
￥6000-6999 51 12.8
￥7000-7999 36 9.0
>=￥8000 22 5.5
Length blog of reading
Under 1 41 10.3
1~2 83 20.8
2~3 142 35.5
3~4 90 22.5
Over 4 44 11
6.3 Validity test of items
In order to measure subjective items of each attribute, several questions of each attribute were proposed, and whether these questions can be combined into a smaller factor which can utmost represent them, factor analysis is required. Kaiser-Meyer-Olkin (KMO) and Bartlett's Test are two tests of factor analysis and they were used to measure the strength of the relationship among variables. Firstly, the KMO measures the sampling adequacy and it should be greater than 0.5, then the factor analysis is meaningful to proceed (Malhotra, Baalbaki and Nasr 2008). Kaiser (1974) pointed out the 0.5 is the minimum level (barely accepted), values between 0.7 and 0.8 are acceptable, and values above 0.9 are superb. Secondly, Bartlett's Test of Sphericity measures the hypothesis that the correlation matrix is an identity matrix, implying variables are uncorrelated. The results can be seen from Table 5, KMO (0.832 > 0.5) shows that factor analysis is appropriate. Also, the null hypothesis of Bartlett's Test of Sphericity test (p-value = 0 < 0.05) is rejected and it means that questions have correlations, and thus it is appropriate to analyze the factors.
Table 5 Kaiser-Meyer-Olkin (KMO) and Bartlett's Test of Sphericity Kaiser-Meyer-Olkin Measure of Sampling Adequacy. 0.832
Bartlett's Test of Sphericity
Approx. Chi-Square 1969.457
df 105
Sig. 0.000
20
eigenvalues (3.61, 2.77 and 2.24) are larger than 1 (Malhotra, Baalbaki and Nasr 2008), which means three factors are proper. Scree plot (Figure 3) also suggests 3 factors are appropriate (the one before the big ―elbow‖). Thus, three factors are extracted and they can accumulatively explain 57.52 % of previous 15 items. In addition, these items can correlate corresponding factors well (see Table 7) because their corresponding factor loadings are all higher than 0.6 after the varimax procedure. Specifically, items ATT 1 to 5 match factor 1, items USE 1 to 5 correlate factor 2, and items NOV 1 to 5 fit in factor 3. This indicates that questions can represent variables‘ information well and the scale is valid.
Table 6 Total variance explained
Componen t
21
Table 7 Correlations between questions
Figure 3 Scree Plot
Corresponding items of NOV 1-5, USE1-5 and ATT 1-5 can be seen in the questionnaire (Appendix 1)
NOV –novelty of content; USE –usefulness of content; ATT –attractiveness of content
6.4 Reliability test
22 Table 8 Reliability test of three variables
Variables N of Items N of cases Cronbach‘s Alpha
NOV 5 400 0.737
USE 5 400 0.785
ATT 5 400 0.888
6.5 Variables’ descriptive statistics and correlations
Seen from Table 9, respondents‘ share intention (meanSI = 3.82, see Table 9) overall is
between neutral (Neither agree nor disagree =3) and agree (agree = 4), and the degree is more inclined to share (close 4). As for the three variables NOV, USE and ATT, their means (3.09, 3.07 and 3.20) are between 3 and 4, but they are more close to 3, which means that respondents‘ attitude tend to neutral. Moreover, the standard deviation of the four variables fluctuate around 1 and it indicates that the volatility of respondents‘ attitude towards these four variables is not so much.
The correlations between independent variables (NOV, USE and ATT) and dependent variable (SI) can be seen from Table 10. All three variables, novelty of content (βNOV=0.244,
p<0.01), usefulness of content (βUSE=0.354, p<0.01) and attractiveness of content
(βATT=0.454, p<0.01) are positively related to share intention under 99%
statistical significance level.
Table 9 Variables‘ descriptive statistics
Variables N Minimum Maximum Mean Std. Deviation
SI 400 1 5 3.82 1.07
NOV 400 1 5 3.09 0.92
USE 400 5 5 3.07 0.96
ATT 400 5 5 3.20 1.13
23
Table 10 Correlations between independent variables and dependent variable
SI NOV USE ATT
SI 1
NOV .244** 1
USE .354** .125* 1
ATT .454** 0.052 0.09 1
Note: n=400.
** Correlation is significant at the 0.01 level (2-tailed) * Correlation is significant at the 0.05 level (2-tailed).
6.6 Hypothesis testing
6.6.1 Manipulation check of conditions
Before testing hypothesis, a manipulation check is needed. Specifically, whether the respondents in the novel (useful, attractive) group indeed perceive the blog as more novel (useful, attractive) or in the less novel (useful, attractive) group indeed perceive it as less novel (useful, attractive) should be confirmed in case of objective manipulation bias. Thus, t-test was used and the null hypothesis is whether the mean s of groups differ (Malhotra, Baalbaki and Nasr 2008). If the p-value is smaller than 0.05, implying that means of groups differ otherwise there is no difference.
24 Table 11 Manipulation check with t-test
Variables Levels mean p-value
NOV Less 2.158 .000 More 3.665 USE Less 2.177 .000 More 3.689 ATT Less 1.783 .000 More 3.812 6.6.2 Regression analysis
In order to explore the relationship between the three independent variables (novelty, usefulness and attractiveness) and dependent variable (share intention) of different conditions, and also the interaction effect of popularity of blogger, two models were constructed (Table 12). Hierarchical regression was used to test the relationships in the model and the values of NOV, USE and ATT are set to 1 (―More‖ = 1) or 0 (―Less‖ = 0) according to their levels. All computations in this section are performed by SPSS. 20
Table 12 Models and variables in the model
Variables included in the model
Model 1
NOV, USE, ATT
Model 2
NOV, USE, ATT ,POP, NOV* POP, USE* POP, ATT* POP
NOV – novelty of content. USE –usefulness of content. ATT –attractiveness of content POP– popularity of blogger
NOV* POP - novelty of content * popularity of blogger USE* POP - usefulness of content * popularity of blogger ATT* POP - attractiveness of content * popularity of blogger
The result (Table 13) shows that both models (Model 1: R2 =0.349, F =67.748, P-value <
0.01; Model 2: R2 =0.350, F= 28.989, P-value < 0.01) reject the null hypothesis that all
parameters of variables are zero, thus both models are feasible.
Moreover, the coefficients of NOV in Model 1 (βNOV1=0.51, p-value < 0.01) and Model 2
25
intention. Also, the coefficients of USE in Model (βUSE1=0.82, p-value < 0.01) and Model 2
(βUSE2 =0.84, p-value < 0.01) indicate that usefulness is positively related to share intention.
Similarly, ATT in Model 1 (βATT1=0.93, p-value < 0.01) and Model 2 (βATT2=0.92, p-value <
0.01) implies a positive relationship between attractiveness with share intention. Therefore, the hypothesis H1, H2 and H3 are supported (Table14). In addition, the parameters of NOV (0.51), USE (0.82) and ATT (0.93) shows that the degree of the influence of attractiveness is higher than usefulness and then novelty (ATT > USE > NOV)
However, the variable popularity of blogger, and its interactions with NOV, USE and ATT (P-value all larger than 0.05) are not significantly related to share intention, so the variable popularity of blogger can not act as a moderator influencing the relationships between NOV, USE, ATT and share intention. In other words, the influence of NOV, USE and ATT on share intention will not be affected no matter whether or not the blogger is popular. Therefore, Hypothesis H4a, H4b and H4c are rejected (Table14).
Table 13 Linear regression
Variable Model1 Model2
B VIF B VIF (Constant) 1.045** 0.99** NOV 0.510** 1 0.580** 2 USE 0.820** 1 0.840** 2 ATT 0.930** 1 0.920** 2 POP 0.110 4 NOV* POP -0.140 3 USE* POP -0.040 3 ATT* POP -0.020 3 R Square 0.391 0.392 Adjusted R Square 0.386 0.381 D-W stat 2.058 2.090 F 84.732** 36.154** NOV–novelty of content. USE –usefulness of content. ATT –attractiveness of content POP– popularity of blogger
26
ATT* POP - attractiveness of content * popularity of blogger ** parameter is significant at the 0.01 level (2-tailed)
*parameter is significant at the 0.05 level (2-tailed).
Table 14 Evaluation of hypothesis
Type of relationship Supported
H1 NOV SI + Yes
H2 USE SI + Yes
H3 ATT SI + Yes
H4a NOV * POP SI — No
H4b USE * POP SI — No
H4c ATT * POP SI + No
7. Discussion and implications
7.1 Summary of results
The research of the article helps understanding the relationships between the content relevance attributes (novelty, usefulness and attractiveness) and share intention, and also helps exploring whether external factor popularity of bloggers will moderate the previous relationships.
The results imply that content relevance attributes indeed play an important role in share intention. All three variables of content relevance, novelty, usefulness and attractiveness have significantly positive effects on share intention and thus support previous hypothesis H1, H2 and H3 (Table 14).
Among these three variables, attractiveness presents more influence than the other two variables on share intention. One possible explanation for the result can be the visual images displayed in the product sharing blog, and these images are always more eye-catching and salient than text and thus it is more easier for people catching sight of images than reading the text (Bird 1992). Also, attractive text is less likely to be rejected by people than less attractive one (Jalalian 2012).
27
Also, the attribute novelty of content is positively correlated to share willingness but the degree is lower than previous two attributes. The reason can be that most products existing in the market barely show real innovations and huge progresses compared to previous products although they are updated all the way. This is because most products have been into a mature and stable stage and people‘s attitudes of products have become more rational, so the influence of novelty will be relatively lower than previous two aspects on share intention. However, contrary to expectation, the moderator popularity of blogger neither influences share intention nor affects the previous three relationships. Specifically, the popularity of bloggers itself is not significantly relevant to share intention and the degree of the correlations between novelty, usefulness, attractiveness and share intention will not be changed no matter how popular the blogger is. Thus hypothesis H4a, H4b and H4c are not supported (Table 14). The finding is opposite to previous studies of endorser effect because reputable endorsers will attract consumers and thus boost firms‘ performance (Deutsch & Ross 2003; Liu, Huang and Minghua 2007). The explanations are presented in the following.
One reason for popularity of blogger not significantly influencing share intention is that reputation is a relative concept (Purohit & Srivastava 2001). Respondents may perceive the blogger in the research setting not so much popular compared with a more reputable blogger identified by respondents themselves, so they may ignore the factor and thus popularity of blogger does not work.
28
attention to popularity of blogger and thus, the reputation factor does not boost the share intention.
7.2 Implications
7.2.1 Implications for research
This study contributes to conceptual learning of the content relevance attributes of sponsored product sharing blogs in affecting share behavior in virtual community and the external factor popularity of blogger was also considered in the model. From theoretical aspect, the result generally implies that content indeed influences people‘s share behavior but popularity of blogger is not an influential factor of share intention.
In the research, content is measured by three content relevant attributes novelty, usefulness and attractiveness in this study. However, content is not just limited to these three attributes, more relevant elements should be considered and included in the model, such as understandability of content, emotion/expressions (emotion icon) or language styles, because these attributes are highly related to individuals‘ preferences or perceptions (Chen, Shang and Li 2014; Berger & Milkman, 2012) and thus may influence share intention. In addition, as the prevalence of online shopping, information searching has become a form of entertainment (Shang, Chen and Shen 2005) thus enjoyment of content can also be considered in the model. Furthermore, more external elements can be added in the model because share behavior is not only influenced by content itself but also by social network ties (Granovetter 1973). Specifically, the tightness of correlations, the amount of time spent and frequency of people interacted with each other are the indicators of social network ties, so these elements may become the external determinants influencing share intention or moderate the relationship between content relevant attributes and share intention, so these elements can be included and explored in the further research.
29 7.2.2 Implications for practice
The results show that the specific three attributes of content novelty, usefulness and attractiveness positively influence people‘s sharing behavior. In real life, consumers are fond of products which could deliver new concept or new feelings. The study conducted by Schoormans and Robben (1997) has indicated that more variations of product packages will attract more consumers‘ attention because consumers care about novel feelings. Thereby, product promotion managers are advised to pay more attention to conveying novel information. In detail, sponsored bloggers should not only recommend new products provided by companies but also display the product in a novel way. For example, when a brand new series of Dior‘s invisible pore foundation recommended in the blog, clear description of the color, texture and its new full-coverage functions could be displayed on oranges instead of the face, because the way to display the effect of the disappearing pores or scratches on the oranges will give consumers a new feelings.
In addition, it is also remarkable to deliver useful product information to consumers, because blogs contain more information (content) volume than advertisement and consumers can know products more clearly before making purchase decisions. Useful information helps make decisions more quickly, which means that consumers can efficiently choose an optimal product and remove inappropriate choices after comparison, so sponsored bloggers are strongly advised to include more product comparison information and summary of product reviews. Thus, ―lazy‖ consumers would be more likely to share the condense information to others and finally purchase especially in such an information overload era.
As for attractiveness, image (logo) in the blog is the most influential element to attract people to read. High quality and aesthetic images, logos or any eye-catching signals are advised to present and match the content. Moreover, the placement of the images or logos should follow matching activation effect (Janiszewski 1988). Specifically, putting images or logos on the right and text on the left can attract more attention because the placement is accord with the way of people‘s hemispheres processing corresponding information.
30
bloggers (opinion leader), because they are not only real consumers of companies but also one of the sources of product information. Their reviews can also be product feedbacks and maybe the future directions of companies making progress. In addition, bloggers who are interested in product sharing should also pay attention to these three aspects of their blog writing and thus increase their popularity.
Furthermore, the findings can also be applied to individual or enterprise webpage design and building. In the research, content shows its importance in virtual community and thus content analysis will be of great significance in webpage design or website building. Specifically, enterprises or individuals who want to increase their website traffic can do a content splitting test with their webpage, which means that market analysts can conduct an experiment containing current webpage (controlled) and few new webpage editions with different changes, such as color, logo or place of ads etc. Then, users are equally distributed into different website editions. After a period of time, when huge differences (website traffic or other figures) between these editions emerge, market analysts collect website metrics such as clicks, amount of sharing and conversion rate. Finally, they can conclude through comparison and determine which edition is the best one.
In recent years, less attention is paid to website content analysis. This is because it is difficult and time-consuming to measure whether these long sentences on webpage is good. Thereby, website splitting test helps solve the relative concept issue and the method is practical and efficient.
31
7.3 Limitations
Although the results are useful and encouraging, some limitations still exist. First of all, only one product category was tested in the experiment, which may result in overgeneralization because different product categories may show diverse share intentions of blogs, especially for gender-related products. For example, lipsticks are favored by female while male use shavers, and then the results of these settings will be different. Such differences although were avoided in the research through adopting a neutral product (a digital camera), interpretation of the results should still be cautious and it is necessary to test different product categories of product sharing blog in the future research.
In addition, the sample is not random since the process of collecting data is through convenient ways, so understanding of findings should be careful. Besides, the sample only includes active respondents who are interested in digital cameras, thus the results may have been influenced by self-selection bias, because those non-active respondents who are also interested in digital cameras may have different opinions toward sharing. Thereby, the results of the research can only be interpreted as share intention of blog by current active camera enthusiasts. Nevertheless, the results are reliable enough because the sample is controlled through the question that asked all the respondents whether they are interested in digital cameras. The respondents whose answers are not interested in digital cameras are eliminated from the sample, because only those people who are interested in the product (the digital camera) give valid answers of the research.
Furthermore, the research only includes limited content relevance attributes in the model. Other content relevant attributes such as understandability of content may also influence respondents‘ share intention and can be included to enrich the model.
32
8. References
Acker, M. and McReynolds, P. (1967). ―The need for novelty: a comparison of six
instruments,‖ The Psychological Record.
Ajzen, I. (1985). ―From Intentions to Actions: A Theory of Planned Behavior,‖
Springer Berlin Heidelberg, (pp. 11-39)
Ajzen, I. (1991). ―The theory of planned behavior,‖ Organizational behavior and
human decision processes, 50(2), pp.179--211.
Andrews, D. (2002). ―Audience-specific online community design,‖ Communications
of the ACM, 45(4), pp.64--68.
Aronson, E., Wilson, T. and Brewer, M. (1998). ―Experimentation in social
psychology,‖ McGraw-Hill, 1, pp.99-142.
Assaker, G., Vinzi, V. and O‘Connor, P. (2011). ―Examining the effect of novelty
seeking, satisfaction, and destination image on tourists‘ return pattern: A two
factor, non-linear latent growth model,‖ Tourism Management, 32(4),
pp.890--901.
Bandura, A. (1986). ―Social foundations of thought and action,‖ 1st ed. Englewood
Cliffs, N.J.: Prentice-Hall.
Bandura, A. (1989). ―Human agency in social cognitive theory,‖ American
psychologist, 44(9), p.1175.
Baxter, M. (1995). Product design. 1st ed. London [u.a.]: Chapman & Hall.
Bello, D. and Etzel, M. (1985). ―The role of novelty in the pleasure travel
experience,‖ Journal of Travel Research, 24(1), pp.20--26.
Berger, J. and Milkman, K. (2012). ―What makes online content viral?‖ Journal of
Marketing Research, 49(2), pp.192--205.
Berlyne, D., Craw, M., Salapatek, P. and Lewis, J. (1963). ―Novelty, complexity,
incongruity, extrinsic motivation, and the GSR,‖ Journal of Experimental
Psychology, 66(6), p.560.
Berlyne, D. (1960). Conflict, arousal, and curiosity. McGraw-Hill Book Company.
Berlyne, D. (1962). Motivational problems raised by exploratory and epistemic
behavior. McGraw-Hill.
Berlyne, D. (1966). Curiosity and exploration. Science, 153(3731), pp.25--33.
Berlyne, D. (1970). ―Novelty, complexity, and hedonic value,‖ Perception &
Psychophysics, 8(5), pp.279--286.
Bock, G. and Kim, Y. (2002). ―Breaking the myths of rewards: An exploratory study
of attitudes about knowledge sharing,‖ Information Resources Management
Journal (IRMJ), 15(2), pp.14--21.
Borlund, P. (2003). ―The concept of relevance in IR,‖ Journal of the American Society
for information Science and Technology, 54(10), pp.913--925.
Bird, L. (1992) ―Eye-catching logos all too often leave fuzzy images in minds of
consumers,‖ (1992). The Wall Street Journal, 220(109), p.B1.
Butler, B., Kiesler, S. and Kraut, R. (2013). ―Community Effort in Online Groups:
Who Does the work and why?‖ Leadership at a distance: Research in
technologically-supported work, p.171.
33
framework for managing knowledge across boundaries,‖ Organization science,
15(5), pp.555--568.
Carson, D. (2008). ―The blogosphere as a market research tool for tourism
destinations: A case study of Australia's Northern Territory,‖ Journal of Vacation
Marketing, 14(2), pp.111-119.
Cattell, R. (1957). Personality and motivation structure and measurement. 1st ed.
Yonkers, N. Y.: World Book Co.
Cattell, R. (1957). Personality and motivation structure and measurement. World
Book Co..
Chen, Y., Shang, R. and Li, M. (2014). ―The effects of perceived relevance of travel
blogs‘ content on the behavioral intention to visit a tourist destination,‖
Computers in Human Behavior, 30, pp.787--799.
Chevalier, J. and Mayzlin, D. (2006). ―The effect of word of mouth on sales: Online
book reviews,‖ Journal of marketing research, 43(3), pp.345--354.
Chiang, I. and Hsieh, C. (2014). ―Exploring the impacts of blog marketing on
consumers,‖ Social Behavior & Personality: An International Journal, 39(9).
Chiou, J., Hsiao, C. and Su, F. (2014). ―Whose online reviews have the most
influences on consumers in cultural offerings? Professional vs. consumer
commentators,‖ Internet Research, 24(3), pp.5--5.
Chiu, C., Hsu, M. and Wang, E. (2006). ―Understanding knowledge sharing in virtual
communities: An integration of social capital and social cognitive theories,‖
Decision support systems, 42(3), pp.1872--1888.
Choi, C. and Scarpa, C. (1992). ―Credible spatial preemption through reputation
extension,‖ International Journal of Industrial Organization, 10(3), pp.439--447.
Chrysochou, P., & Krystallis, A. (2013). Understanding consumer-generate
advertising for mainstream fashion goods.
http://works.bepress.com/chrysochou/45/Chu, S. and Kamal, S. (2008). ―The effect of perceived blogger credibility and
argument quality on message elaboration and brand attitudes: an exploratory
study,‖ Journal of Interactive Advertising, 8(2), pp.26--37.
Chu, S. and Kim, Y. (2011). ―Determinants of consumer engagement in electronic
word-of-mouth (eWOM) in social networking sites,‖ International Journal of
Advertising, 30(1), pp.47--75.
Compeau, D. and Higgins, C. (1995). ―Computer self-efficacy: Development of a
measure and initial test,‖ MIS quarterly, (19).
Cosijn, E. and Ingwersen, P. (2000). ―Dimensions of relevance,‖ Information
Processing & Management, 36(4), pp.533--550.
Crompton, J. L. (1979). ―Motivations for pleasure vacation,‖ Annals of tourism
research, 6(4), 408-424.
Crozier, W. (1994). Manufactured pleasures. 1st ed. Manchester: Manchester
University Press.、
Cupchik, G. (1999). Emotion and industrial design: reconciling meanings and
feelings.
Proceedings of the 1st International Conference on Design and
Emotion
. pp.75-82.
34
