Exploring individual communication power in the blogosphere
The Authors
Hsiu-Chia Ko, Department of Information Management, Chaoyang University of Technology, Wufeng, Taiwan, Republic of China
Chun-Po Yin, Department of Data Processing, Haiching Vocational High School of Technology and Commerce, Kaohsiung, Taiwan, Republic of China
Feng-Yang Kuo, Department of Information Management, National Sun Yat-Sen University, Kaohsiung, Taiwan, Republic of China
Abstract
Purpose – Viewing the blog technology as an integral part of the current social-technical environment, this research aims to investigate whether the main influences on message diffusion within a blog community originate from external mass media channels or internal interpersonal communication channels.
Design/methodology/approach – The authors employed the innovation-diffusion model and the OLS estimating method to study message diffusion of two documentary films on the Wretch, the largest blog community in Taiwan.
Findings – The results indicate that the mass media is the main source of message diffusion and that the internal communication power may increase as the opinion leader promotes these messages.
Research limitations/implications – Other factors that may influence message diffusion such as topic, design characteristics, and the existing social network have not been included.
Practical implications – For practice, the result indicates that the mass media and the blog might complement each other.
Originality/value – This research is one of the first that attempts to apply the innovation-diffusion model to analyze message diffusion within the blog community.
Article Type:
Research paper
Keyword(s):
Worldwide web; Communication; Social networks.
Journal:
Internet Research
Volume:
18
Number:
5
Year:
2008
pp:
541-561
Copyright ©
Emerald Group Publishing Limited
ISSN:
1066-2243
1 Introduction
The blog is becoming one of the most popular communication channels on the Internet. The Technorati report of February 2006 indicates that a new blog is created every second, and that there were 60 times more blogs in 2006 than in 2003.This rapid increase in the number of blogs has been seen almost everywhere, including Taiwan where many Internet service providers such as MSN Space, Sina, Xuite, the Wretch blog web site (www.wretch.cc), Sky, and RoDo are offering such services to potential users.
Consequently, many experts now consider that blogs may be changing the way that messages are communicated via the mass media. Traditionally, the populace has passively received messages from media institutions such as newspaper and television. On the other hand, in blogging, messages are published and diffused in blog communities based on the interconnected nature of bloggers and the facilities provided by these communities, such as blogrolls, web site recommendations, citations, linkages, TrackBack, Really Simple Syndication, and the exchange of opinions between friends (Marlow, 2004; Lee et al., 2006). Given that audiences can create their own news platform and communication channels, a considerable shift in how we make – and consume – the news has taken place (Gillmor, 2004). For example, Lasica (2001) considered the blogging phenomenon to be a grassroots movement that may sow the seeds for new forms of journalism, public discourse, interactivity, and online communities. Also, Johnson and Kaye (2004) found that bloggers believed blogs to be more credible than traditional media.
Thus, there exist important considerations that need to be answered: will blogs replace traditional media such as newspaper; will they complement each other because both are integral parts of today's social-technical world, or will blogs simply become attached to the mainstream mass media and thereby be subject to the same constraints thereon? In this study, we have addressed these issues using the theory of innovation diffusion. According to Rogers (1995), diffusion is the process by which an innovation is communicated through certain channels over time among the members of a social system. Rogers theorizes that mass media channels (outside the social system) and interpersonal channels (inside the social system) represent two modes of message diffusion that influence whether people assimilate a message and decide to adopt an innovation (Rogers, 1995). Previously, researchers have applied the innovation-diffusion theory to analyze news events, technological innovation, and new communication technologies (Rogers, 2003). Jensen (2002) has pointed out that innovation-diffusion theory can be applied to both the diffusion of new media technology and message diffusion within media. Since blogs are considered to be a type of new communication technology, in this study we have applied the innovation-diffusion theory to investigate how messages diffuse in a blog community. Specifically, we have investigated whether the main influences on message diffusion within a blog community originate from external mass media channels or internal interpersonal communication channels through two case studies. We have also explored the effect of the diffusion of messages within a blog community that are associated with many related reports in the mass media. By tracing the diffusion processes of two cases within a blog community using the time series of posts, this research not only aids our understanding of the communication power of blogs, but also provides information that is useful to general communication research and applications.
2 Literature review
2.1 Introduction to blog
Herring and Wright (2004) defined a blog as frequently modified web pages in which dated entries were listed in reverse chronological sequence. The contents and forms of blogs are rich and diverse. A blog can be a personal diary, a memo, a place to deliver messages, a cooperation field, a news-release web site, or a temporary performing stage for a politician. Blood (2002) distinguished three basic types of blogs:
- filters;
- personal journals; and
- notebooks.
The content of filters is external to the blogger (e.g., world events and online happenings), while the content of personal journals is internal (the thoughts and internal workings of the blogger); notebooks may contain both, and are distinguished by longer, more-focused essays.
According to Blood (2002), filter blogs were initially the preeminent type of blog. Now, however, the majority of blogs are of the personal journal variety. Currently, a blog may be not only a place for people to document their life or communicate with relatives and friends, but also a news source. Some people treat a blog as a channel on which to present their opinions on various topics. The articles of some bloggers have even been collected as published physical books.
Blogging is a relatively new phenomenon, most of the applied research methods have been exploratory, and the research issues have also varied. One interesting area of study is related to the power exerted by individuals who publish through blogs. It has been shown that blogs can significantly influence politics such as the resignation of Trent Lott (Drezner and Farrell, 2004; Gillmor, 2004), the delivering of first-hand information about the Iraq War (Gillmor, 2004), and the coverage of the Tulip Revolution (Kulikova and Perlmutter, 2007). This increased ability of individuals and groups to publish messages and handle sources of information on the Internet means that media institutions no longer have a centralized power to control in controlling the power of communication. Bloggers can be both publishers and critics of news report, and opinions of bloggers can also represent a source for related news reports.
Investigations into the motivations of blogging (Nardi et al., 2004; Schiano et al., 2004) show that in bloggers document their life, provide commentary and opinions, express opinions to influence others, think by writing, release emotional tension, perform a social activity, articulate ideas through writing, and form and maintain community forums. Efimova's study (2003) reveals that the motivations of bloggers include curiosity, improving ones' information management and learning abilities, and a general interest in sharing ideas with others. In addition, some researchers investigate the reasons why bloggers continue to share/post information on their blogs (Lu and Hsiao, 2007; Miura, 2007; Miura and Yamashita, 2007). For example, Lu and Hsiao (2007) have found that personal outcome expectations such as to enhance self-image and gain praise have stronger influence on the information sharing behavior. Similarly, Trammell and Keshelashvili's (2005) content analysis finds that popular bloggers often actively engage in impression management by demonstrating competence through presentation of their intellectual abilities. Furthermore, gender may play an interesting role in blogs, as men are more likely to maintain the traditional filter blog whereas women are more likely to run a diary/journal blog (Pedersen and Macafee, 2007). For example, the study by Harp and Tremayne (2006) has examined gender inequity among the most-read political blogs based on network theory and feminist theory. Their findings show that, similar to opinion writing and the publishing world, women's voices are marginalized in the blogsphere, probably due to the fact that men are more active in politics than women and outnumber women on the web. Finally, Viegas (2005) finds that some bloggers are acutely aware that they do not know their audiences, and cannot control who accesses their blogs, indicating that privacy is an important issue of blogging.
In summary, the past research has explored how and why individuals adopt blogs, continues to use blog, and issues such as gender equity and privacy. Yet rarely have researchers focused on understanding the impact of individual communication power on blogging behavior. This research is one the first attempts to apply the innovation-diffusion model to analyze message diffusion within the blog community. By viewing the blog technology as an integral part of the current social-technical environment, we have employed the innovation-diffusion model and the OLS estimating method to study message diffusion of two documentary films on Wretch, the largest blog community in Taiwan, in this study. This analysis allows us to investigate which media, i.e. Wretch or the mass media, exerts higher influence on message diffusion within Wretch.
2.2 Diffusion of innovation
The introduction of the diffusion concept by Rogers has led to a large body of literature. According to the definition of Rogers (1995), “an innovation is an idea, practice, or object that is perceived as new by an individual or other unit of adoption”. The perceived newness of the idea determines the individual's reaction to it: “Diffusion is the process by which an innovation is communicated through certain channels over time among the members of a social system” (Rogers, 1995). This definition reflects that diffusion is a special type of communication, in that the messages are about a new idea.
According to the definition, there are four main elements associated with the diffusion of innovations:
- the innovation;
- communication channels;
- time; and
- the social system (Rogers, 1995).
Furthermore, there are three basic diffusion models:
- internal influences (word-of-mouth (WOM) or interpersonal communication);
- external influences (mass media communication); and
- mixed influences (Loh and Venkatraman, 1992; Mahajan et al., 1990; Rogers, 1995; Venkatraman et al., 1994).
2.2.1 Internal-influence model
Mansfield (1961) proposed the internal-influence model, in which diffusion occurs through channels of communication within a social system and is appropriate for testing the imitation hypothesis. Applied to our study, a blogger writes an article related to the same topic in the blog community that may be based on other bloggers. The diffusion derives from imitative behavior within the social system. The model can be stated as: Equation 1 where N(t) is the cumulative number of adopters at time t, m is the total number of people who will eventually adopt the innovation, and q is the coefficient of internal influence. In this model, the diffusion rate is a function of the number of people who have already adopted the innovation and the remaining number of potential adopters. The model is shown graphically in Figure 1.
2.2.2 External-influence model
The external-influence model assumes that the rate of diffusion at time t is dependent only on the potential number of adopters present at that time. Information about the innovation is available from a source external to the social system, and the model does not consider interaction between previous adopters and potential adopters. In our study, the sources of external influence were the mass media, including newspapers and TV. The model can be stated as: Equation 2 where p is the coefficient of external influence and is a nonnegative constant, and N(t) results in a curve that increases at a decreasing rate (Figure 2).
2.2.3 Mixed-influence model
The mixed-influence model subsumes both the internal- and external-influence models (Bass, 1969). The mixed-influence model can be stated as: Equation 3 The cumulative distribution of the model appears as an S-shaped generalized logistic curve whose shape depends on coefficients p and q. The model is the most general and is widely used to combine the internal and external influences. The diffusion of the model combines coverage by the media and the early adoption.
2.2.4 Summary
The essence of the diffusion process is the information exchange through which one individual communicates a new idea to one or more other individuals. In studies-related message diffusion, one issue interesting to researchers concerns how people become aware of “events” via either mass media or personal channels (McQuail, 2002). Research focusing on tracing the spread of a spectacular news event has shown that the distribution of knowers over time follows an S-shaped curve, with interpersonal and mass media channels playing comparable roles (Cohen, 2002; Fine, 1975; Funkhouser and McCombs, 1971; Gantz et al., 1986; Jeffres and Quarles, 1983; Mayer et al., 1990; Riffe and Stovall, 1989; Rogers, 2000; Rogers, 2003; Smythe, 2002). For example, Riffe and Stovall (1989) have investigated the role of emotional response on the diffusion of shuttle disaster-related messages. They found that the cumulative percentages of aware time of the respondents exhibited the familiar steep diffusion S-curve, and there were no significantly emotional response between people who learned the explosion via mass media and via interpersonal communication. In another study, Cohen (2002) examined the social functions fulfilled by Israel's broadcast media following the assassination of Prime Minister Itzhak Rabin. He found that both television and radio were important initial sources about the news of the assassination, and television was important for following subsequent events. Finally, the meta-review on past news event diffusion research by Rogers (2000) concluded that early studies have focused on the S-shaped rate of diffusion over time, media and interpersonal channels, and the salience of news events.
Recently, Rogers and Seidel (2002) found that TV was faster than the Internet for the immediate diffusion of news about the 911 terrorist attacks. The news of the attack presented on TV was perceived as very salient, and diffused rapidly. Individuals reacted to the news in an emotional way and relied on TV to obtain instantaneous information. Person-to-person communication was also important in diffusing awareness-knowledge of the terrorist attacks, but the Internet became more important later when people sought in-depth information. Thus, it appears that a news item may have a powerful but short media exposure due to rapid loss of its novelty value and being replaced by other news events. This contrasts with a blog on the Internet, which can be a platform for recurring items related to a particular news event. A blog has a continuous drumbeat that builds in volume and cadence as more information is added to it (Gill, 2004).
In this study, we treat the publishing and diffusion of messages (such as news events) within a blog community as a type of innovation diffusion, in which the community members have different perceptions of the degree of newness or the attraction, and the messages diffuse in the blog community through the links and widespread audience of a blog (i.e. an interpersonal communication). The contributions of each blogger have time stamps, and these contributions can be collected according to the adoption rates of innovations related to the same issues. The diffusion of messages in a blog is influenced by both the internal social system (the connection of blogs) and the external mass media channels (e.g., newspaper and TV). The innovation-diffusion theory was therefore adopted here to study message diffusion within the blog community.
3 Research questions and methodology
This study addressed two research questions:
RQ1. Which channel, interpersonal or mass media, provides the main influence on the message diffusion in a blog community?
RQ2. How does the external mass media influence the message diffusion in a blog community?
Specifically, we aimed to understand whether the interpersonal communication power in a blog community would be influenced by the presence of a large number of news items related to the same topics being reported in the mass media. Furthermore, because the typical diffusion curve had an S shape, the diffusion patterns might differ between before and after the inflection point. We treated the week with the most news reports appeared in the mass media as an inflection point, which was used to divide the entire diffusion period into two distinct data sets. We then analyzed the patterns and explanatory power for changes between before and after the infection point. This methodology allowed us to determine whether the number of articles in a blog community increased due to the large quantities of related news items reported by the mass media.
3.1 Data collection method
We selected the Wretch blog platform (www.wretch.cc), which has the largest number of members in Taiwan (Yam, 2005), as our research sample. Also, according to Insightxplorer (2005), from December 2004 to February 2005 Wretch experienced an average visitor rate (the percentage of all Internet users to visit a specific web site) of 10 percent. In March 2005, Wretch had more than 400,000 registered members, and was growing at up to 2,000 members daily. Finally, the data flow of Wretch was second only to Yahoo in Taiwan (Chen, 2006). Thus, we considered it to be a suitable research target.
The next step was to explore the patterns of message diffusion in the Wretch community. We chose the following two documentary films as samples of our case study for exploring the communication power of Wretch:
- Chronicle of the Sea, Nan-Fang-Ao (CSNFA).
- Let It Be (LIB).
There were several reasons for choosing these two films for our study. First, both documentary films were noncommercial and fell outside the mainstream movie industry in Taiwan. Their promotion fee was far lower than that for commercial movies and, as a result, the blog platform became an important channel for movie fans to learn about the two films. Next, their subjects were similar: LIB was a documentary about rice farmers and their land, while CSNFA told a story about fishermen and the sea. Both films emphasized how farmers/fishermen struggled with their lives. Finally, both documentary films had a similar process of exposure to the public media: LIB received mass media attention when it was televised on Public Television on September 30, 2004, and when it received an award at the Taiwan International Documentary Festival on November 14, 2004; similarly, CSNFA gained media attention when it received an award at the Taipei Gold Horse film festival on December 5, 2004. The mass media exposure of both LIB and CSNFA peaked in May 2005, when they were released in theaters. Data on LIB were collected over 60 weeks from September 30, 2004, until November 23, 2005, while data on CSNFA were collected over 51 weeks from November 4, 2004, until October 26, 2005.
The following procedure was applied to collect the study data. First, LIB and CSNFA were used as Wretch keywords to obtain a list of the IDs and URLs of bloggers who posted articles related to the films. A computer program was written to automatically record this list and extract the IDs and URLs. Another computer program was written to capture the contents of all articles from the URLs, and to extract the published time-stamp data for each article and the blogroll lists for each ID. Finally, two data sets were obtained: the first contained the time series data of the articles that were used to analyze the process of message diffusion, and the second contained the relationships of IDs (based on the blogroll linkages) that were used for social network analysis. Tables I and II list the publication frequencies of the two films according to the publication date, and Figures 3 and 4 show the accumulative curves for the two data sets.
Note that while the data collection periods for the two films overlapped each other substantially, they attracted greatly differing levels of interest on Wretch and received different levels of mass media attention during this time. The contrasting patterns of discussions on the blog and attentions from mass media about the two films allowed us to compare their information diffusion patterns, and investigate how the mass media channel affects the information diffusion patterns in the Wretch community.
To elucidate the effects of external influences (mass media channels) and internal influences (interpersonal channels in a blog), we relied on news reports about LIB and CSNFA from UDN and Chinatimes, and used the data to investigate the relationship between the Wretch and the mass media. Funkhouser and McCombs (1971) considered news events to be the easiest type of information in which to study diffusion, since they have definite starting points in time and are generally easy to comprehend and define. For this study, we choose the UDN electronic knowledge database and Chinatimes electronic news database because together they represented the largest database for news reports. The UDN was established by the UDN Group at the beginning of 2001, since when it has collected news reports from all the newspapers in the database. It has the largest number of readers in Taiwan. The Chinatimes database was established by Chinatimes Group, and has collected more than 320 million news reports since 1994. Moreover, in Taiwan, essentially the same news events were reported in all TV channels and newspapers. Thus, news reports about the two films could serve as the basis of our analysis. Tables III and IV list the sum of numbers of news articles related to both documentary films in UDN and Chinatimes.
3.2 Data analysis framework
The model of Bass (1969) assumes that potential adopters of an innovation are influenced by two means of communication:
- the mass media; and
- WOM.
This further assumes that the adopters of an innovation comprise two groups:
- those influenced by mass media communication (an external influence); and
- those influenced by interpersonal communication (an internal influence) (Mahajan et al., 1990).
Mahajan et al. (1988) studied the adoption of the M-form organization structure using regression-analogue equations (ordinary least squares) to test different diffusion models. The analysis method used in the present study was based on Bass (1969) and Mahajan et al. (1988).
3.3 Research procedures
3.3.1 The null hypothesis
A stringent null hypothesis would assume that the diffusion pattern follows a white-noise or random-walk process (Loh and Venkatraman, 1992; Mahajan et al., 1988). The white-noise process implies that the difference between the numbers of adopters at t and (t−1) is random. That is, in time-series data, the rate of diffusion is driven only by the error term: Equation 4 where x(t) is the number of adopters at time t, and the residuals ɛ(t) have a zero mean that is uncorrelated with ɛ(t−k) for all nonzero k. The model represented by a time series will proceed by a sequence of unconnected steps, starting each time from the previous value of the adopted time series.
3.3.2 External-influence model
The regression analogue of the external-influence model, equation (2), for describing a time-series adoption pattern can be stated as (Mahajan et al., 1988): Equation 5 where x(t−1)=N(t)−N(t−1) is the number of adopters at time t, β 2=(1−p) with β 2 <1, and the residuals, ɛ(t), have zero mean and are uncorrelated with ɛ(t−k) for k≠0. If β 2=1, equation (5) reduces to the white-noise model.
3.3.3 Mixed-influence model
Two popular innovation-diffusion models that generate the S-shaped adoption curve and especially capture imitation behavior are those proposed by Mansfield (1961) and Bass (1969): Equation 6 where the second term on the right-hand side represents the contact between adopters and nonadopters, thus reflecting imitation behavior. Consequently, the nonnegative constant q is usually defined as the coefficient of internal influence. Note that q=0 in equation (6) yields the Coleman (external-influence) model (i.e. Equation (3)).
One possible regression analogue of both imitational (mixed and internal) models can be derived as: Equation 7 where β 2=1 + q−p for the Bass (mixed-influence) model and β 2=1 + q for the Mansfield (internal-influence) model, with β 2>1, β 3<0, and N *(t−1)= N 2(t−1)−N 2(t−2) (Mahajan et al., 1988).
3.3.4 Internal-influence model
In mixed-influence equation (7), β 2=1 + q−p for the Bass (mixed-influence) model and β 2=1 + q for the Mansfield (internal-influence) model, with β 2>1 and β 3<0. In external-influence equation (5), β 2=(1−p) and β 2 <1. So, we derive the internal-influence β 2 as 1+(mixed-influence β 2)−(external-influence β 2): Equation 8 where β 2=1+(1 + q−p)−(1−p), with β 2>1 and β 3<0.
Finally, we employed the F-test (Mahajan et al., 1988; Venkatraman et al., 1994) to test the significance of the explanatory power of alternative models. The F-test was also adopted for comparing the white-noise model with the three alternative models.
3.3.5 Evaluating the two research questions
Table V lists the four models and their parameters. For RQ1 (Section 3), we evaluated the main source of diffusion based on the results of analyses of these models. For RQ2, we treated the week with the highest number of related news items as the inflection point, and estimated the diffusion patterns before and after this week to determine whether the diffusion patterns were affected by reports in the mass media of related news. Tables III and IV demonstrate that LIB and CSNFA had the highest numbers of news (mass media) and blog entries in the 34th and 31st weeks, respectively. Subsequently, we analyzed the variations in the sources of influence in the two periods.
4 Results of data analysis
4.1 Correlation analysis
Figures 5 and 6 show the message-diffusion patterns for the two films as graphs of time-series data for the number of news items in the mass media and the number of articles on the Wretch. These figures suggest that the number of news items and the number of blog articles were positively correlated.
We next applied a correlation analysis to quantify the relationship. Table VI summarizes the analysis results, which indicates that the number of articles on the Wretch web site was positively correlated with the number of news items in the mass media for both films. However, we cannot verify whether the power of autonomy is present in the blog writing; this is addressed in the next section based on analysis results for the different diffusion models.
4.2 Analysis results of diffusion models
The results of the F-test based on the entire diffusion period of the two films are listed in Tables VII and VIII. For LIB, the external and mixed models were significant relative to the white-noise model, whereas for CSNFA, all three alternative models were all significant relative to the white-noise model. These results indicate that most of the alternative models were better fits than the white-noise model for the two films.
Tables IX and X summarize the fitting statistics (adjusted R 2 values) of the three alternative models and the white-noise (random) model for LIB and CSNFA, respectively, (the details of the parameter estimations are given in Appendix A). To answer the RQ1 – the main influence on message diffusion in a blog community – the analysis results show that for LIB, the external model (adjusted R 2=0.5740, p=0.0459) provides a better fit to the diffusion pattern, indicating that the external mass media had the greatest influence.
For CSNFA (Table X), although the mixed model (adjusted R 2=0.4802, p=0.0216) better fits for the diffusion pattern, the mixed model combines the internal and external influences, and the explanatory power of the external model (adjusted R 2=0.4720) is higher than that of the internal model (adjusted R 2=0.4583). Thus, the result indicates that the main influence is also the external mass media. The external model of the two films shows higher explanatory power than the internal model. Therefore, the mass media provides more influence on the message diffusion than the blogs' internal social network in the Wretch for both of the cases considered.
For the RQ2, the analysis results of Section 4.1 show the correlation between the blog community and the mass media. We now consider the diffusion process before and after the inflection point. For LIB, before the 34th week, the internal- and mixed-influence models had the same explanatory power (adjusted R 2=0.8346). It is clear that before the inflection point (the 34th week when LIB was released), the internal communication channels inside the blog community had a greater influence than the external communication channels (adjusted R 2=0.8346 > 0.5953). However, after the films were released, the external communication channels had a greater influence than internal communication channels inside the blog community (adjusted R 2=0.6680 > 0.3245). The diffusion patterns differed between before and after the 34th week (the inflection point), thus suggesting that the quantity of news reports influenced the message-diffusion process on the Wretch for this case. For CSNFA, the analysis results show that the external model demonstrates a higher explanatory power before and after the inflection point (31st week).
Comparing the diffusion patterns of LIB and CSNFA revealed that the internal communication power was higher for the former than for the latter before the inflection point. The method of social network analysis was further applied to analyze the patterns of the message-diffusion networks of the two films. The relationship sets from all blogroll lists linked to other blogs represented the bloggers' relationships, and were transformed into a directed matrix (Garton et al., 1997), whose rows and columns were the initiators and receivers of the linkage relations, respectively. Table XI depicts a part of the matrix, in which the value of “1” for a cell indicates the existence of a relationship, while the value of “0” indicates no relationship. The number of the title of column and row is the serial number of blogger in the Wretch, and can correspond to the social network graphic (Figure 7). Netminer software was used to transform the relationship matrix into the complete social network. The complete social networks for the two films at the end of the data collection period are shown in Figures 7 and 8.
The results indicate that there were two opinion leaders during the message-diffusion process of LIB. The first leader's blog (first in Figure 7) contained many in-degree connections with other blogs, meaning that many bloggers regularly visited his/her blog and read his/her articles. The blog of the other opinion leader (second in Figure 7) incorporated a LIB category that documented all reports about this film, and this leader also had many in-degree connections. However, the diffusion network of CSNFA displayed a relatively loose pattern and only one relationship (Figure 8) in the entire network. These two different patterns seem to confirm the suggestion by Rogers (2003) who considers that opinion leaders play important roles in diffusion networks. This is consistent with the findings of several studies that have shown the greatest influence within the blog community may come from a handful of elite bloggers who are associated with most of in-degree connections (Adamic and Glance, 2005; Drezner and Farrell, 2004).
5 Discussion and conclusions
In this study, we investigated the process of message diffusion through two case studies in the blog community, and whether mass communication channels or interpersonal communication channels have a greater influence. There are numerous optimistic reports about the communication power of blogs, which consider them to be a publishing tool for individuals. Our research results based on the innovation-diffusion model show that the communication power is greater for the mass media than for the interpersonal communications of the blog community observed in this study. In other words, blogs in present-day Taiwan do not yet have sufficient autonomy or internal power of communication, and they were unable to diffuse messages efficiently in the two cases considered. This may be due to the fact that the majority of blogs in Taiwan are diaries (Insightxplorer, 2005). As this type of blogs are used by bloggers mainly for the purpose of sharing their lives with close friends rather than attracting public attention, their internal communication power would be less influential than the external communication power of the mass media in the diffusion of blog messages.
In addition, when tracing the first case, our study shows that the social network of opinion leaders can enhance the level of message diffusion within a blog community. Thus, while the research results demonstrate that the mass media exerts a high level of influence on the communication and transmission of messages at present, the internal communication power of a blog may increase as the opinion leader promotes these messages. Several studies have shown that blogs specializing in political issues exhibit a power-law distribution (Adamic and Glance, 2005; Drezner and Farrell, 2004). That is, the small number of elite bloggers who are associated with most of in-degree connections may have the greatest influence on the message diffusion within the blog social network. Hence, the organizations who want to apply the interpersonal communication power of blogs to disseminate the message automatically and effectively, they have to consider how to find the opinion leaders.
5.1 Implications for research
While some scholars suggest that blogging phenomenon to be a grassroots movement that may sow the seeds for new forms of journalism, public discourse, interactivity, and online communities (Lasica, 2001), our research shows that blogs are integral part of today's social-technical world and cannot be treated independently of other media. As blogs and other Web 2.0 applications like Facebook are growing at present, the line to delineate public and private domains also becomes more and blur. Our analysis of the first case in this study indicates that the importance of the mass media and inter-blog communication may change over time, with the inter-blog communication exerting higher influence at the beginning (before the 34th week in Table IX). It is likely that the continuing increases in the prevalence of blogs and other Web 2.0 applications may result in increased internal communication power in the world of Web 2.0, especially when individuals interconnect themselves and form a strong social network. It is therefore crucial for researchers to investigate the power of communication for various media in the era of Web 2.0 in the future.
5.2 Implications for practice
The blog has been advocated as a powerful interpersonal communication channel through its WOM ability. Our research result shows that while the mass media has dominant power during the message diffusion process, the first case analysis demonstrates that, through the opinion leaders, inter-blog social network may enhance blogs' communication power. This indicates that the mass media and the blog should be viewed as integral parts of today's social-technical world. Gill (2004) has considered that news item have a powerful but short media exposure due to rapid loss of its novelty value, but blogs have a continuous drumbeat to build in volume and cadence as more information is added to it. Thus, if businesses and organizations are trying to find ways of utilizing blog for purposes such as marketing and advertising, they should combine mass media and build a strong social network for the blog's opinion leaders.
5.3 Limitations and future research
This study was subject to a few limitations. First, we only investigated the diffusion of documentary films through two case studies, and diffusion patterns of other topics such as politics and entertainment may differ. Second, we only considered Wretch blog platform, and other blog communities may exhibit different characteristics. Third, it is difficult to collect complete news reports from the numerous types of mass media in Taiwan, and our use of the UDN and Chinatimes electronic news database may have introduced bias. Fourth, the design of the blog platform and the characteristics of the blog community (e.g., the demographic distribution of members or the type of blogs) may also affect the structure of a social network and the diffusion of messages. The effects of all of these factors should be explored in future studies. Finally, past studies have shown that social factors such as interpersonal communication, social situation, and social structure can affect the diffusion of news events and messages (Fine, 1975; Rogers, 2003). This study was conducted in Taiwan, and the result might differ if conducted elsewhere since the social factors will differ in other countries. Hence, generalizing the results of this study to other countries or cultures should be carefully conducted.
Equation 1
Equation 2
Equation 3
Equation 4
Equation 5
Equation 6
Equation 7
Equation 8
Figure 1Internal influence diffusion curve
Figure 2External-influence diffusion curve
Figure 3Accumulated number of discussion articles related to LIB
Figure 4Accumulated number of discussion articles related to CSNFA
Figure 5News in the mass media and articles in blogs related to LIB
Figure 6News in the mass media and articles in blogs related to CSNFA
Figure 7Social network of LIB at the end of the data collection period
Figure 8Social network of CSNFA at the end of the data collection period
Table IDiscussion articles related to LIB
Table IIDiscussion articles related to CSNFA
Table IIINews articles related to LIB
Table IVNews articles related to CSNFA
Table VSummary of hypothesized model specifications
Table VICorrelation analysis
Table VIIComparison of the diffusion models with the white-noise model of LIB
Table VIIIComparison of the diffusion models with the white-noise model of CSNFA
Table IXAnalysis of diffusion models of LIB
Table XAnalysis of diffusion models of CSNFA
Table XIA part of the relationship matrix in this study
Table AIParameter estimation of LIB diffusion
Table AIIParameter estimation of LIB diffusion before the 34th week
Table AIII
Table AIVParameter estimation of CSNFA diffusion
Table AVParameter estimation of CSNFA diffusion before the 31st week
Table AVIParameter estimation of CSNFA diffusion after the 31st week
References
Adamic, L., Glance, N. (2005), "The political blogosphere and the 2004 US election: divided they blog", mimeograph, .
Bass, F.M. (1969), "A new product growth for model consumer durables", Management Science, Vol. 15 pp.215-27.
Blood, R. (2002), "Introduction", in Rodzvilla, J. (Eds),We've Got Blog, The Perseus Books Group, Cambridge, MA, .
Chen, S. (2006), "The Wretch", Common Wealth, Vol. 356 pp.164-5.
Cohen, Y. (2002), "Broadcast news diffusion in crisis-ridden democracies: Israel and the Rabin assassination", Harvard International Journal of Press/Politics, Vol. 7 pp.14-33.
Drezner, D.W., Farrell, H. (2004), The Power and Politics of Blogs, American Political Science Association, Washington, DC, .
Efimova, L. (2003), "Blogs: the stickiness factor", BlogTalk: A European Conference on Weblogs, Vienna, .
Fine, G.A. (1975), "Recall of information about diffusion of major news event", Journalism Quarterly, Vol. 52 pp.751-5.
Funkhouser, G.R., McCombs, M.E. (1971), "The rise and fall of news diffusion", Public Opinion Quarterly, Vol. 35 pp.107-13.
Gantz, W., Krendl, K.A., Robertson, S.R. (1986), "Diffusion of a proximate news event", Journalism Quarterly, Vol. 63 pp.282-7.
Garton, L., Haythornthwaite, C., Wellman, B. (1997), "Studying online social networks", Journal of Computer Mediated Communication, pp.3.
Gill, K.E. (2004), "How can we measure the influence of the blogosphere", WWW2004, New York, NY, .
Gillmor, D. (2004), We the Media, O'Reilly, O'Reilly Media, Inc., Sebastopol, CA, .
Harp, D., Tremayne, M. (2006), "The gendered blogosphere: examining inequality using network and feminist theory", Journalism and Mass Communication Quarterly, Vol. 83 pp.247-64.
Herring, S.C., Wright, E. (2004), "Bridging the gap: a genre analysis of weblogs", paper presented at 37th Hawaii International Conference on System Sciences, .
Insightxplorer (2005), "Blog continue to be hot and popular", A Report of ARO Measurement of Network, Insightxplorer, Taipei, .
Jeffres, L.W., Quarles, R. (1983), "A panel study of news diffusion", Journalism Quarterly, Vol. 60 pp.722-5.
Jensen, K.B. (2002), A Handbook of Media and Communication Research: Qualitative and Quantitative Methodologies, Routledge, London, .
Johnson, T.J., Kaye, B.K. (2004), "Wag the blog: how reliance on traditional media and the internet influence credibility perceptions of weblogs among blog users", Journalism and Mass Communication Quarterly, Vol. 81 pp.622-42.
Kulikova, S.V., Perlmutter, D.D. (2007), "Blogging down the dictator? The Kyrgyz revolution and Samizdat websites", The International Communication Gazette, Vol. 69 pp.29-50.
Lasica, J.D. (2001), "Blogging as a form of journalism", Online Journalism Review, .
Lee, M.K.O., Lim, C.K.H., Sia, C.L. (2006), "Understanding customer knowledge sharing in web-based discussion boards", Internet Research, Vol. 16 pp.1066-2243.
Loh, L., Venkatraman, N. (1992), "Diffusion of information technology outsourcing: influence sources and the Kodak effect", Information Systems Research, Vol. 3 pp.334-58.
Lu, H., Hsiao, K. (2007), "Understanding intention to continuously share information on weblogs", Internet Research, Vol. 17 pp.345-61.
McQuail, D. (2002), McQuail's Reader in Mass Communication Theory, Sage, London, .
Mahajan, V., Peterson, R. (1985), Models for Innovation Diffusion, Beverly Hills, CA, Sage, .
Mahajan, V., Sharma, S., Bettis, R.A. (1988), "The adoption of the M-form organizational structure: a test of imitation hypothesis", Management Science, Vol. 34 pp.1188-201.
Mahajan, V., Muller, E., Bass, F.M. (1990), "New product diffusion model in marketing: a review and directions for research", Journal of Marketing, Vol. 54 pp.1-26.
Mansfield, E. (1961), "Technical change and the rate of imitation", Econometrica, Vol. 29 pp.741-66.
Marlow, C. (2004), "Audience, structure and authority in the weblog community", paper presented at International Communication Association Conference, New Orleans, LA, .
Mayer, M.E., Gudykunst, W.B., Perrill, N.K., Merrill, B.D. (1990), "A comparison of competing models of the news diffusion process", Western Journal of Speech Communication, Vol. 54 pp.113-23.
Miura, A. (2007), "Can weblogs cause the emergence of social intelligence? Causal model of intention to continue publishing weblog in Japan", AI & Society, Vol. 22 pp.237-51.
Miura, A., Yamashita, K. (2007), "Psychological and social influences on blog writing: an online survey of blog authors in Japan", Journal of Computer-Mediated Communication, pp.12.
Nardi, B.A., Schiano, D.J., Gumbrecht, M., Swartz, L. (2004), "Why we blog", Communications of the ACM, Vol. 47 pp.41-6.
Pedersen, S., Macafee, C. (2007), "Gender differences in British blogging", Journal of Computer-Mediated Communication, pp.12.
Riffe, D., Stovall, J.G. (1989), "Diffusion of news of shuttle disaster: what role for emotional response?", Journalism Quarterly, Vol. 66 pp.551-6.
Rogers, E.M. (1995), Diffusion of Innovation, The Free Press, New York, NY, .
Rogers, E.M. (2000), "Reflection on news event diffusion research", Journalism and Mass Communication Quarterly, Vol. 77 pp.561-76.
Rogers, E.M. (2003), Diffusion of Innovation, The Free Press, New York, NY, .
Rogers, E.M., Seidel, N. (2002), "Diffusion of news of the terrorist attacks of September 11", Prometheus, Vol. 20 pp.209-19.
Schiano, D.J., Nardi, B.A., Gumbrecht, M., Swartz, L. (2004), "Blogging by the rest of us", Stanford University White Papers, pp.12.
Smythe, T.C. (2002), "The diffusion of the urban daily", Journalism History, Vol. 28 pp.73.
Trammell, K.D., Keshelashvili, A. (2005), "Examining the new influencers: a self-presentation study of A-list blogs", Journalism and Mass Communication Quarterly, Vol. 82 pp.968-82.
Venkatraman, N., Loh, L., Koh, J. (1994), "The adoption of corporate governance mechanisms: a test of competing diffusion models", Management Science, Vol. 40 pp.496-507.
Viegas, F.B. (2005), "Bloggers' expectations of privacy and accountability: an initial survey", Journal of Computer-Mediated Communication, pp.10.
Yam (2005), A Survey of Internet Users in Taiwan, Yam, .
Appendix. Parameter estimation for the four models of the two films considered
Corresponding author
Feng-Yang Kuo can be contacted at: bkuo@mis.nsysu.edu.tw