WJMCR 26 (November 2010)
Recently, the arrest of “Minerva” has been one of the most fervently debated subjects online and offline throughout South Korea. A person known as “Minerva”, the username for Daum Agora, was recently arrested for writing about the worsening global economy and criticisms about current Korean economic policies. Believing him to be a digital opinion leader, the government charged him with the crime of electronically spreading false information that damaged the public good. This case has recently re-captured international attention, suggesting a need for research concerning online opinion leadership in the new media environment.
Despite the enduring popularity of the concept of opinion leadership, academic researchers have only recently begun to explore the ramifications of online opinion leadership. A review of the literature in this area reveals that recent studies of online opinion leaders are plagued by inconsistencies in operational definitions and clear delineations of the attributes of online opinion leaders. In addition, existing research has focused almost exclusively on opinion leaders, neglecting examination of non-leaders and other questions to be answered.
To extend opinion leadership theory into new theoretical domains and settings beyond its traditional emphases, this study applied the concept of opinion leadership to the Internet environment. A content analysis of messages posted to an online message board was utilized to examine how various participant statuses within the continuum of opinion leadership (e.g., opinion leaders, quiet persuaders, general public) use different types (e.g. Interactive, reactive, other) and categories of messages (e.g., fact, opinion, question, other).
The picture emerging from the findings indicate that online opinion leaders as well as other participant groups differ from those identified and characterized in traditional studies of opinion leadership. Results show that both online opinion leaders and non-leaders are more similar than they are different in terms of the interactivity and messages that they post. All posters, regardless of opinion leader status, most often wrote what Rafaeli would classify as “reactive messages” to express their opinion. The popularity of reactive messages challenges the general assumption that interactive messages are the most desired form of communication in the online, networked environment.
Despite widely held assumptions about the virtues of interactive message flows, the findings of this analysis suggest that “reactive” messages might constitute enough interaction to maintain a meaningful thread or conversation. Yet, the popularity of reactive messages and the reasons for the predominance of this type of message require further research for a more thorough explanation. What is notable is that general public was not necessarily writing messages to seek information, challenging the assumption that media audiences are primarily information and opinion seekers. Taken together, the findings of this study suggest that the online environment does not necessarily encourage interactive message exchanges among discussants; however, it does appear to promote mutual discourse and role interchanges among participants of a news-oriented online community.
Recently, the arrest of “Minerva” has been one of the most fervently debated subjects online and offline throughout South Korea.1 A person known as “Minerva”, the username for Daum Agora,2 was recently arrested for writing about the worsening global economy and criticisms about current Korean economic policies. Believing him to be a digital opinion leader, the government charged him with the crime of electronically spreading false information that damaged the public good. This case has recently re-captured international attention, suggesting a need for research concerning online opinion leadership in the new media environment.
Despite the enduring popularity of the concept of opinion leadership, academic researchers have only recently begun to explore the ramifications of online opinion leadership. A review of the literature in this area reveals that recent studies of online opinion leaders are plagued by inconsistencies in operational definitions and clear delineations of the attributes of online opinion leaders. In addition, existing research has focused almost exclusively on opinion leaders, neglecting examination of non-leaders and other questions to be answered. To extend opinion leadership theory into new theoretical domains and settings beyond its traditional emphases, this study applied the concept of opinion leadership to the Internet environment.
With the advent of new media, particularly the Internet, the term, “interactivity,” has been increasingly used as a feature of the technologies. The idea is that characteristics of the different technologies render different degrees of “interactive” communications. Therefore, “new” media3 are associated with new technology and high interactivity. On the contrary, traditional, “old” media are associated with old technology and low interactivity. It is assumed that users of new interactive media have “better” experience. This technological deterministic view is popular and prevalent, and arguments are made that increased interactivity will result in substantial benefits in areas ranging from business to politics.
Unfortunately, the complexity of the concept of interactivity has driven divergent approaches in explicating what interactivity is. Today, interactivity is often defined and studied from three different perspectives: definitions that focus on ‘features of medium’, focus on ‘processes’, and lastly concentrating on ‘user’s perception.’4 Despite the long period of efforts to reach consensus on definition, scholars seem to diverge on conceptualizing interactivity under different frameworks. However, this is not necessarily problematic because there are different dimensions of the concept. This paper will approach interactivity from communication process-related perspective defined by Rafaeli.
According to Rafaeli, interactivity resides within continuous two-way interactions among communicators: “a variable characteristic of communication settings… an expression of the extent that in a given series of communication exchanges, any third (or later) transmission (or message) is related to the degree to which previous exchanges referred to even earlier transmissions.”5 Under this perspective, Rafaeli’s message-centered responsiveness model conceptualizes interactivity as meaning-relatedness that is constructed through two-way communication between the sender and receiver. When a later message does not recount the relatedness of earlier messages, it is demoted to the second level of communication called reactive communication, creating ambiguous levels of interactivity. The first level of interactivity would be one way communication where there is no chain of interrelated messages at all. Unfortunately, Rafaeli does not provide a clear demonstration of what the interactive message exchange ought to be, making his conceptualization incomplete and difficult to employ.
Besides the lack of a concrete example of just what is an interactive message exchange to be, Rafaeli’s conceptualization of interactivity is opaque in that it equates the exchange of messages with the exchange of meaning, according to Newhagen.6 This is an erroneous assumption because there is no established empirical research to support such implication. In fact, Rafaeli’s own study conducted with Sudweeks7 demonstrates just the opposite conclusion: there is an inverse correlation between number of message exchanges and amount of meaning-relatedness. Nevertheless, this assumption is inevitable because Rafaeli bases messages- meaning relatedness for the conceptualization of interactivity while he uses the exchange of messages to operationalize it. In this logic, Newhagen notes Rafaeli’s categorization of message exchanges is more appropriate for operationalizing the concept of interactivity in communication rather than using it as a theoretical framework.8 Following Newhagen’s logic, this paper will use Rafaeli’s categorization of message exchanges for operationalizing the level of interactivity of individual messages.
Heavily focused on the quality of two-way message flows, this process-related approach lacks understanding of media effects or the outcomes of these message exchanges on users.9 Nevertheless, Schultz attempted to apply Rafaeli’s responsiveness model to examine the effects of interaction between New York Times journalists and readers via e-mails and online forums10. He hypothesized that exchanges of messages between readers and journalists would narrow the knowledge gap among them and increase public’s political participation. Although his predictions were not supported, Schultz’s study is significant in the way it tried to investigate the areas that the responsiveness model normally omits, the role exchange of communicators.
Schultz’s study is meaningful in that it brings back the idea of role exchange into the responsive model. For many years, some scholars have insisted that the egalitarian notion of mutual discourse and role exchange in two-way communication is what defines interactivity.11 Pavlik later developed the definition of interactivity to include multi-directional communication; he argues, “Interactivity means two-way communication between source and receiver, or, more broadly multidirectional communication between any number of sources and receivers.”12 In the study, he suggests two categories of interactivity when considering journalistic websites: reader to reader and journalist to reader interactivity. These websites allowed multi-directional communication but serve more to inform users than to form mutual discourse, encouraging more one-way communication in nature. Although Schultz did not state this, it is possible to speculate that the one-way nature of communication is likely to create vertical flow of communication from journalists to readers, upgrading the status of journalists as opinion leaders while downgrading readers as followers. The concept of vertical flow and opinion leadership were originally developed from the two-step flow model.
Introduced in an era dominated by theories of powerful media effects, the interrelated ideas of opinion leaders and the two-step flow model are well known for writing the new paradigm for the limited effects of media. Although the model was not originally formulated in the original voting study conducted by Lazarsfeld, Berelson and Gaudet in 1948, this particular study is generally viewed as the first draft for the initial model.13 Subsequently, the model was developed after further studies were conducted based on the conclusions of the original study. In assessing the results of the research, the authors discovered the influence of personal contact, particularly person-to-person, in the process of voting behavior, shedding light on the idea of a two-step model involving opinion leaders. They observed, “ideas often flow from radio and print to opinion leaders and from these to the less active sections of the public.”14
Starting from Lazarsfeld’s panel survey to Noelle-Neumann’s ‘personality strength (PS),’15 the identification of opinion leadership has been a primary concern in off-line opinion leadership research. As Katz himself criticized, previous studies on offline opinion leaders were overly occupied with the demographics of opinion leaders and their socio-political characteristics as well as individual differences of those who are considered ‘opinion leaders.’16 Recently, research interest in opinion leadership has expanded its domains to identifying dispositional characteristics,17 adding social embeddings of opinion leaders,18 and examining the model in different settings where not only socio-cultural environment but also the media environment differs.19 The effort to incorporate the evolving media environment has only received preliminary research attention.20
Given the social diffusion of interactive media technologies, the interrelated concept of the two-step flow model and opinion leadership may require some reconsideration. In this spirit, this paper attempts to apply the concepts of opinion leadership to interactive media environment under the framework of Rafaeli’s responsiveness model of interactivity. Currently, Rafaeli and Sudweeks’s study in 1997 appears to be the only precedent research that examined the idea of online opinion leadership under the responsiveness model. In the study, they identified opinion leaders as discussants who wrote messages most frequently. Assuming that type of messages by frequent authors differ in nature of interactivity from those of infrequent authors, they hypothesized, “Interactivity is related to individual activity and communication salience of participants.” Accordingly, it was found that messages written by frequent (active) discussants were significantly more reactive than others (a reactive message is a message that responds to one preceding message). Although this paper uses different measurements for classifying discussion participants, it expects to observe similar results. Thus, the following hypothesis can be proposed:
H1: Participants who are identified as opinion leaders will most frequently write messages that respond to at least one preceding message, in other words reactive messages.
Despite the fact that Rafaeli and Sudweeks study21 is one of the seminal researches in this area, its limited scope of analysis on opinion leadership and relationship with interactivity bound its applicability to public opinion research. Attempting to develop the existing research, this paper investigates messages of opinion leaders as well as those of followers under the framework of responsiveness model. In line with the work of Rafaeli and Sudweeks, an exploratory research question is posed:
RQ1: How is message interactivity related to one’s status within the continuum of opinion leadership?
RQ2: How are contents of messages related to one’s status within the continuum of opinion leadership?
Beyond the relationship between opinion leadership and interactivity, one’s status within a discussion forum or online community the opinion leadership is also related to the content of messages that one posts. According to previous studies, the willingness to transmit information is a primary characteristic of opinion leaders.22 Based on this finding, this present study intends to compare the characteristics of traditional opinion leaders on online and “online opinion leaders” in terms of the types of message contents (i.e. fact, opinion, question, and others) in which both types of opinion leaders use. By comparing the message types, this present study can test previous role exchange ideas of interactivity. Therefore, the present study adds a variable, message types, to the existing literature on opinion leadership and messages.
H2: Participants who are identified as opinion leaders will be the most likely to write messages to transmit factual information or knowledge.
Because the research questions are intended to examine the interactive message exchanges related to opinion leadership, content analysis is used for the present study. Since the purpose of this paper is to analyze text-based exchanges using the responsiveness model, a content analysis of messages posted to an online message board was utilized to examine how various participant statuses within the continuum of opinion leadership (e.g., opinion leaders, quiet persuaders, general public) use different types (e.g. interactive, reactive, other23) and categories of messages (e.g., fact, opinion, question, other).
Among different formats of interactive online activities, the political discussion board was selected for analyses. The discussion board, by nature, provides a relatively free-flowing discussion environment where moderators are not needed to lead participants in a discussion of the issues. In such a free environment which encourages many interactions among participants, the choices and preferences of participants form certain social roles. Thus, discussion boards are an opportune place to observe opinion leaders, attention gatherers, and general public.
Because this present research required many messages, popularity was the key criterion for selecting a discussion board. Popularity of a discussion board can be measured through search engine rankings. Therefore, a simple search was conducted via the Google search engine. The search terms were “politics”, “discussion”, AND “messageboards”. As a result, ‘UsmessageBoard,’24which was at the top of the list, was chosen. This message board fit the purpose of the present study in that it is operated by a non-governmental commercial party which provides space for active discussions for every Internet user who wants to participate for free. At the same time, the message board was one of the largest active sites in terms of the number of subscribers and posts.
Out of its various categories of forums, “Healthcare/Insurance/Govt Healthcare” was selected for the sampling frame. These forums should not be considered as mere information boards, but arenas of discourse where people share their opinion as well as knowledge. The messages for the study sample were randomly selected from all the messages written between June 1, 2009 to August 30, 2009. The origination date for this research was chosen based on The New York Times Healthcare timeline.25 According to this timeline, June 1, 2009 was the first day of the summer congressional debate and August 30, 2009 was the final day. Then, composite day random sampling procedure was used to avoid possible weekend bias and to ensure that messages from all days were represented equally in the study sample. In total, the study sample included 14 days over the three-month period.26
This study relied upon a content analysis approach to examine the types and nature of messages that different groups of discussants wrote. Thus, data collection for this research entailed the categorization of discussants. Instead of dividing discussants into a strict dichotomy of opinion leaders and followers, this study categorized discussants within a continuum of opinion leadership by including a group called “Attention gatherers” situated between opinion leaders and the general public. Thus, there are three categories of discussant groups: (1) opinion leaders, (2) attention gatherers and (3) general public.27
As for the categorization of participants, two scale measurements (Thanks and Reputation power) that represent one’s status as a discussant were employed. Both systems are derived from the online activities data that were objectively collected by the host site. Both “Thanks” and “Reputation power” are cumulative systems which add the number of points that individual discussants earn respectively from other discussants (“Thanks”) and from the administrator (“Reputation power”). However, for the purpose of this study, only discussants’ points gained by “Thanks” are analyzed. As mentioned, “Thanks” points are directly given by discussant whereas “Reputation power” is an indirect way of gaining status points because administrator intervene between discussants’ interactions. Unlike “Reputation power”, “Thanks” points cannot be changed by the system. The message board readily shows who gives Thanks to whom to discussants and the public. Therefore, “Thanks” system reflects the true views of discussants.
This study employed the classification system used by Rhee, Kim and Kim in 200728 to categorize discussion participants and identify online opinion leaders. It assigned specific number value to each group of the three categories of discussant groups: opinion leaders (10%), attention gatherers (10%), and the general public (80%). Therefore, this present research assumed that similar percentages of the discussant population also can be identified with each group in UsmessageBoard. In the sample of this research, “Thanks” points are the percentage of posts that receive thanks.
When this value was larger than 20%, the discussant was identified as an opinion leader, which represent about 10% of the sample. When this value was between 10% and 19%, the discussant was categorized as an attention gatherer, which represents about 10% of the sample. When this value was between 0% and 9%, the discussant was classified as a general public, which represents about 80% of the sample. Due to the lack of precedent studies which identify each category of discussant by using Thanks system, this present study adopted and modified the scheme that Rhee, Kim and Kim produced.
As seen in Figure 1, individual messages were categorized into three different participator type (opinion leader, attention gatherer, general public). Next, level of message interactivity of each participatory type were identified and coded as interactive, reactive, or “other” type of messages following the Rafaeli’s responsiveness model. The fundamental definitions for the message kind were as follows: (1) Interactive message: a message containing references to the manner in which previous messages related to those preceding them (2) Reactive message: a message that responds to one preceding message (3) Others: a message that does not fall into any of the above categories. Lastly, message type of each message interactivity is identified and coded as fact, opinion, question, and “other” nature of message.
By using UsmessageBoard, this present research is able to capture the technological influence on user-to-user interactivity. UsmessageBoard has a particular feature, called quote. As seen in Figure 2, discussion participants can capture a part or a whole message in which other participants have written; then, they can incorporate the captured quote into their own messages. For example as seen in Figure 2, KittenKoder quotes DavidS, who quotes KittenKoder. This quoting can continue indefinitely. This feature of the message board is implemented for users’ interactivity, making a thread more thorough and related. Thus, this study analyzed site-specific techniques to determine the exact nature of message-based interactivity. In short, present study examined opinion dynamics in an online message forum determined by technological as well as non-technological factors. Finally, all the messages were again categorized as to whether they were facts, opinions, questions, or other categories.
All the findings are based on descriptive statistics. For this study, 15 participants (10%) were classified as opinion leaders via the Thanks system according to the definition advanced earlier. A total of 15 posters (10%) were identified as attention gatherers while the rest, 126 participants (80%), were considered as the general public.
By and large, all three types of interactivity were evenly distributed. However, reactive messages were the most often used for discussions by all posters (40.1%) while interactive messages comprised 38.7% of the messages. A relatively large number of messages (21.1%) was identified as “others” level of message interactivity. Specifically, 40.8% of the messages written by opinion leaders were reactive messages. Therefore, Hypothesis 1 received limited support because the difference in percentage between reactive and interactive messages was not substantial. Similarly, the general public had also used reactive messages the most (39.7%). Yet, their use of interactive messages was not much greater than reactive messages: 38.2% of messages were interactive. Lastly, attention gatherers wrote 41.8% of interactive messages and 41.8 % of the reactive messages.
Hypothesis 2 was unsupported. Instead, opinion leaders posted opinionated messages the most (63.5%). Messages primarily concerning facts (11.4%) and questions (7.6%) together comprised 19% of total messages while other topics occupied 17.5% of the messages that opinion leaders wrote. In fact, all participants, regardless of status, were most likely to post opinionated messages: Far higher attention gatherer, this number was (62.4%) while for the general public infers the number was even (67.1%). The second most prevalent type of content belonged to the other category: attention gatherers (19.9%) and general public (15.4%). Messages categorized as others for its content involved compliments, personal attacks, and jokes. Following after the category of “others”, the category of fact was placed as the third most dominant written type of content (10.4%): attention gatherers (11.3%) and general public (9.5%). Lastly, question occupied 7.8% of the entire messages: attention gatherers (6.4%) and general public (8.1%).
Considering both the nature of interactivity and type of content, opinion leaders most often used reactive messages to convey their opinions (65.1%). Both attention gatherers (64.4%) and general public (72.7%) relied heavily on reactive messages. Opinion leaders (16.7%) and general public (19.9%) most frequently chose non-interactive messages to distribute facts, whereas interactive messages were the most popular way of disseminating facts for attention gatherers. To ask or pose a question, both opinion leaders (9.3%) and general public posters (8.2%) mostly preferred to write reactive messages when attention gatherers continued to prefer the interactive messages. Finally, both attention gatherers and general public most often wrote interactive messages for ‘other’ type of contents while opinion leaders wrote reactive messages for the similar purposes.
As a preliminary analysis, this study examined the types and nature of messages that opinion leaders and followers posted to a popular political discussion forum. The picture emerging from the findings indicates that online opinion leaders as well as other participant groups differ from those who are identified and characterized in traditional studies of opinion leadership. Results show that both online opinion leaders and followers are more similar than they are different in terms of the interactivity and nature of messages that they post.
Although the data did not support Hypothesis 2, this preliminary analysis points to the need for methodological improvements and more nuanced analyses. Methodologically, the present analysis systematically identified opinion leaders and other participant groups. Previous studies largely overlooked the influence of individual messages when categorizing discussion participants. On the other hand, this present study attempted to capture the influence of individual messages by incorporating a Thanks point system. Furthermore, this research overcame the limitation of Rafaeli’s definition of interactivity by adding technological attribute– i.e. quotes feature in figure 2- as another determining factor of interactivity.
Even though this study received limited empirical supports, one noteworthy finding of this preliminary research is that all posters, regardless of their status within the continuum of opinion leadership, most often wrote what Rafaeli would classify as “reactive messages” (40.1%) to express their opinions. The popularity of reactive messages challenges our general assumption that interactive messages are the most desirable and effective way of communicating. Despite a longstanding reverence of interactivity, the finding suggests that reactive messages might be enough interaction to maintain a meaningful thread or conversation. Yet, the popularity of reactive messages requires further investigation for more concrete explanations. Online opinion leaders seem to use reactive messages the most. The finding suggests that it would be wrong to classify one’s status within the continuum of online opinion leadership by their use of message interactivity. Thus, there are grounds to question the general expectation that online opinion leaders should take the initiative in adapting to interactive environments by writing interactive messages.
If discussion forum participants are relatively similar in terms of the interactivity that they use, then another factor must be driving opinion leadership. The similarity of different participant groups can be found in terms of the contents that they use; it was found that all discussants had written opinionated messages by far the most. Considering the purpose of the message boards, this particular result is not a noteworthy finding. What is notable is that general public group was not necessarily writing messages to seek information, challenging the general assumption that they are information and opinion seekers. Rather, the findings indicate that non-opinion leaders play on active role in information flow. Although opinion leaders had a leading position in disseminating information to others, only the small percentage of messages were written for the purpose of transmitting information (11.4 %). In the meantime, 9.5% of messages written by the general public had the purpose of distributing information. Only emphasizing opinion leaders’ role in the information flow blurs the characteristics of traditional opinion leaders and others.
Taken together, the findings of this study suggest that the online environment does not necessarily encourage interactive messages among discussants, however, it does seem to promote mutual discourse and role interchanges among the discussants.
Despite the popularity of the concept of opinion leadership, current studies about online opinion leaders are hampered by pre-Internet definitions of opinion leadership and their attributes, leaving other, perhaps more important, questions unaddressed. Only recently, have academic researchers begun to explore the world of online opinion leadership.
Apparently, there are areas, perhaps, more important than identification and attribute searches, that need to be investigated. In this respect, the present analysis attempted to expand the analytical context, observing the interplay among interactive messages, message types, and participant status within a continuum of opinion leadership. Despite the difficulty of classifying discussion participants, the picture emerging from the findings hints that online opinion leaders as well as other participant groups differ from those identified and characterized in traditional studies of opinion leadership. Specifically, the findings indicate that opinion leaders do not necessarily prefer writing interactive messages nor do they solely write to transmit knowledge and facts to others. In this respect, some of the general assumptions about opinion leadership should probably be revisited.
Although the present study acknowledges the importance of outcomes or consequences associated with interactive message flows, Rafaeli’s definition of interactivity is limiting. Following Rafaeli’s definition of interactivity with few technological features of the system only allowed conclusions to be drawn that specifically related to message interactivity. In this respect, more research is needed on the possible consequences of the relationship between interactivity and opinion leadership.
Nevertheless, this analysis has much to contribute to our understanding of online opinion leadership and interactivity. This paper explored interactivity as a multi-dimensional construct and incorporates the concept of opinion leadership that combines the dimensions of participant’s control and direction of communication flows. Despite limited support for the hypotheses, the results imply that interactivity should be approached as a multi-dimensional construct rather than from a single definitional perspective. On this note, the present analysis contributes to the established literature by providing empirical data where it is still rarely collected.
Ji won Kim is a graduate student in telecommunications at Indiana University. Joonseok Choi was a graduate student in telecommunications at Indiana when this paper was written and is now a doctoral student at the University of Texas.
Classification by ‘Thanks’ & Message Interactivity
|Thanks * Message Interactivity Cross tabulation|
|Thanks||Opinion leaders= Over 20 %||Count||83||86||42||211|
|% within Thanks||39.3%||40.8%||19.9%||100.0%|
|% within Interactivity||14.4%||14.4%||13.3%||14.1%|
|% of Total||5.6%||5.8%||2.8%||14.1%|
|Attention gatherers= 10%-19%||Count||59||59||23||141|
|% within Thanks||41.8%||41.8%||16.3%||100.0%|
|% within Interactivity||10.2%||9.9%||7.3%||9.5%|
|% of Total||4.0%||4.0%||1.5%||9.5%|
|% within Thanks||38.2%||39.7%||22.0%||100.0%|
|% within Interactivity||75.4%||75.8%||79.4%||76.4%|
|% of Total||29.2%||30.4%||16.8%||76.4%|
|% within Thanks||38.7%||40.1%||21.2%||100.0%|
|% within Interactivity||100.0%||100.0%||100.0%||100.0%|
|% of Total||38.7%||40.1%||21.2%||100.0%|
Classification by ‘Thanks’ & Message Type
|Thanks * Message Type Cross tabulation|
|Thanks||Opinion leaders= Over 20 %||Count||24||134||16||37||211|
|% within Thanks||11.4%||63.5%||7.6%||17.5%||100.0%|
|% within content||16.2%||13.6%||13.7%||15.4%||14.1%|
|% of Total||1.6%||9.0%||1.1%||2.5%||14.1%|
|Attention gatherers= 10%-19%||Count||16||88||9||28||141|
|% within Thanks||11.3%||62.4%||6.4%||19.9%||100.0%|
|% within content||10.8%||8.9%||7.7%||11.7%||9.5%|
|% of Total||1.1%||5.9%||.6%||1.9%||9.5%|
|% within Thanks||9.5%||67.1%||8.1%||15.4%||100.0%|
|% within content||73.0%||77.5%||78.6%||72.9%||76.4%|
|% of Total||7.2%||51.3%||6.2%||11.7%||76.4%|
|% within Thanks||9.9%||66.2%||7.8%||16.1%||100.0%|
|% within content||100.0%||100.0%||100.0%||100.0%||100.0%|
|% of Total||9.9%||66.2%||7.8%||16.1%||100.0%|
Opinion leader: Message Interactivity and Message Type
|Message Interactivity * Message Type * Thanks Cross tabulation|
|Opinion leaders= Over 20 %||Interactivity||Count||10||54||6||13||83|
|% within content||41.7%||40.3%||37.5%||35.1%||39.3%|
|% of Total||4.7%||25.6%||2.8%||6.2%||39.3%|
|% within content||29.2%||39.6%||50.0%||48.6%||40.8%|
|% of Total||3.3%||25.1%||3.8%||8.5%||40.8%|
|% within content||29.2%||20.1%||12.5%||16.2%||19.9%|
|% of Total||3.3%||12.8%||.9%||2.8%||19.9%|
|% within content||100.0%||100.0%||100.0%||100.0%||100.0%|
|% of Total||11.4%||63.5%||7.6%||17.5%||100.0%|
Attention Gatherer: Message Interactivity and Message Type
|Message Interactivity * Message Type * Thanks Cross tabulation|
|Attention gatherers= 10%-19%||Interactivity||Count||7||34||5||13||59|
|% within content||43.8%||38.6%||55.6%||46.4%||41.8%|
|% of Total||5.0%||24.1%||3.5%||9.2%||41.8%|
|% within content||31.3%||43.2%||44.4%||42.9%||41.8%|
|% of Total||3.5%||27.0%||2.8%||8.5%||41.8%|
|% within content||25.0%||18.2%||.0%||10.7%||16.3%|
|% of Total||2.8%||11.3%||.0%||2.1%||16.3%|
|% within content||100.0%||100.0%||100.0%||100.0%||100.0%|
|% of Total||11.3%||62.4%||6.4%||19.9%||100.0%|
General Public: Message Interactivity and Message Type
|Message Interactivity * Message Type * Thanks Cross tabulation|
|General public= 0%-9%||Interactivity||Count||28||282||32||94||436|
|% within content||25.9%||36.9%||34.8%||53.7%||38.2%|
|% of Total||2.5%||24.7%||2.8%||8.2%||38.2%|
|% within content||27.8%||42.7%||40.2%||33.7%||39.7%|
|% of Total||2.6%||28.7%||3.2%||5.2%||39.7%|
|% within Interactivity||19.9%||62.2%||9.2%||8.8%||100.0%|
|% within content||46.3%||20.4%||25.0%||12.6%||22.0%|
|% of Total||4.4%||13.7%||2.0%||1.9%||22.0%|
|% within content||100.0%||100.0%||100.0%||100.0%||100.0%|
|% of Total||9.5%||67.1%||8.1%||15.4%||100.0%|
CODING SHEET 1: THE MESSAGE
Classification of Discussion Participants
Division of discussants by ‘Thanks’
(Total number of ‘Thanks’ received / Number of total Posts * 100)
|Opinion Leader||Attention Gatherers||General Public|
|Thanks (%)||Over 20%||10% – 19%||0% – 9%|
Nature of Interactivity
|1||Interactive messages||A message containing references to the manner in which previous messages related to those preceding them|
|2||Reactive message||A message that is responding to one or more preceding messages|
|3||Others||Does not fall into any of the above categories|
Content of the Messages
|1||Fact||A message primarily contains a fact|
|2||Opinion||A message primarily contains an opinion|
|3||Question||A message primarily contains a question/request|
|4||Others||A message that does not fall into any of the above categories|
Note: Adapted and modified from Rafaeli & Sudweeks, 1997
CODING SHEET 2: THE MESSAGE
- Interactive message: A message containing references to the manner in which previous messages related to those preceding them: Mn(Mn-1 (Mn-K-1) ). It includes messages that contained number referenced e.g., “In # 28 message”, and quoting. Quoting includes various ways: (1) cutting and pasting selections of previous messages; (2) using software features to quote previous messages e.g, “subject:Re” or “Jason writes:”; (3) referring explicitly to contents of previous messages e.g.,“ In your message..”.
- Reactive message: A message responds to one preceding message: Mn (Mn-1) . It does not require a message to contain references to the manner in which previous message related to those preceding them. But it simply responds to the question, information, or opinionated statements that were discussed in preceding messages.
- Others: A message that does not fall into any of the above categories
- Fact: A message primarily contains description or presentation of the fact that is known as truth or something known to have happened. That is, information communicated or received concerning a particular fact or circumstance such as news e.g. “In the speech that the president Obama gave…” It includes youtube videos of broadcasts and news articles but not blog posts that has any opinions in it. In this sense, opinionated TV programs or news articles as well as content that contains sarcasms, criticism, and caricature cannot be identified as facts. Arguments about the fact counts as a fact.
- Opinion: A message primarily contains expressions of personal thoughts and views such as agreement/disagreement with others or content of other’s messages e.g., “I do not see the way as you do.” Opinionated TV programs or news articles as well as contents that contain sarcasms, criticism, and caricature are identified as opinion. Also,stories on personal experiences or anecdotes are opinions as well.
- Question: A message primarily contains requests for information or questions for clarifications. It includes messages asking for explanations for the definitions of technical terms and for the procedure/history of the phenomena.
- Others: A message’s nature that does not fall into any of the above categories. Personal questions (e.g., self-revealing questions; “where do you live?”) or statements that arouse humor fall under the category of others. This includes irrelevant statements (i.e., Off-topic random statements) and statements that make personal attack on someone else. Opinions about the specific person in the message board fall under this category.
SAMPLE FRAME: COMPOSITE DAY
|June 1 2009||(Monday)||Number of Messages: 1|
|June 9 2009||(Tuesday)||Number of Messages: 88|
|June 17 2009||(Wednesday)||Number of Messages: 113|
|June 25 2009||(Thursday)||Number of Messages: 29|
|July 3 2009||(Friday)||Number of Messages: 121|
|July 11 2009||(Saturday)||Number of Messages: 28|
|July 12 2009||(Sunday)||Number of Messages: 49|
|July 20 2009||(Monday)||Number of Messages: 52|
|July 28 2009||(Tuesday)||Number of Messages: 140|
|August 5 2009||(Wednesday)||Number of Messages: 54|
|August 13 2009||(Thursday)||Number of Messages: 152|
|August 21 2009||(Friday)||Number of Messages: 177|
|August 29 2009||(Saturday)||Number of Messages: 304|
|August 30 2009||(Sunday)||Number of Messages: 184|