Literature Review
There is a small but growing literature on people using computer mediated communication
(CMC) tools, in general and to discuss politics. Important themes in this literature include the
ability of computer mediated communication (CMC) to support the creation of communities or at
least personal relationships, the characteristics of CMC that seem to encourage negative comments
and uninhibited behavior, and the reasons that people participate in CMC at all. These themes are
present in the literature that looks at CMC in general and in the research that has looked specifically
at political discussion online.
These online communities have some characteristics that should attract the researcher. Much of the discussion in these communities is carried out in public for anyone who is interested to read. Many groups are automatically archived by DejaNews (http://www.dejanews.com), an online search engines, which makes messages available from any of the archived groups through a keyword search. As with real world communities, of course, not all discussion is available to the researcher. Internet etiquette (netiquette) implies that flames and angry arguments should be conducted by private email so that the whole group is not interrupted. Purely social messages might be considered off topic in many groups, and so they would be carried on via private email as well. Posters can form personal relationships with people that they meet online and these relationships often spill over into face to face meetings perhaps to add sensory information not available online, (Parks and Floyd, 1996) especially for those users that post often and over an extended period of time. (Walther , 1992; Walther, Anderson, and Park, 1994) "Rather than engendering loneliness and isolation, they may well serve to link people. Rather than removing societal norms, they may well be a vehicle for establishing them." (Rafaeli, 1986, p. 125) Rheingold (1995) built a picture of the development of strong intimate relationships online with his anecdotal description of behavior on The Well. People who "visit" several usenet groups on a similar interest will tend to see some familiar users and that strengthens their relationship. (Baym, 1996; McLaughlin, Osborne, and Smith,1995; Wellman and Gulia, 1996) Participants develop rules of behavior for particular groups; posters are encouraged to "act" online in accordance with these rules and reminded when they display proscribed behavior. People should not send personal attacks to the whole group but rather send them to through email to that specific person. (MacKinnon, 1996) Another is to not type in all capital letters except when "shouting". The rules are not written down however and they are not the same from group to group. Politically oriented newsgroups have characteristics similar to real world political groups; they discuss politics, post information about their group or cause in order to attract new members, and they work to maintain the group's dominant political ideology. (Hill and Hughes,1997) These communities are based on shared interests, rather than shared social characteristics (Wellman and Gulia, 1996; McLaughlin, Osborne & Smith, 1996, p. 91) Participants do more than just exchange information. They offer support to each other, give opinions, share jokes. They organize real world activities. (Rash, 1997; Wellman and Gulia, 1996) Through interaction over time they can develop a sense of trust, sustained commitments, shared values, and some significant sense of caring or mutual obligation, characteristics of members of a community (Kling, 1996).
There are some problems with considering CMC groups as communities, especially with politically oriented CMC. (Bimber, 1997) Unlike most real world communities and personal interactions, it is very easy to join and to leave these communities. People can belong to many online communities simultaneously, which may limit the attention and concern they can pay to each. (Jones, 1995; Wellman and Gulia, 1996) Posters do not know exactly who will read their words in these online community or who will respond. It takes time for misunderstandings to be discovered and corrected. (MacKinnon, 1996) Archived messages and lurkers can result in a loss of confidentiality and anonymity, which can be a threat to the community. People to whom you were not speaking directly have the capacity to get an exact copy of your words, months even years after the original message was posted. (Kling, 1996) Interaction over time is no guarantee of civil behavior either. Frequent posters on political newsgroups used their messages to tear other posters down, to dominate the discussion, and to drive out dissenting opinion. (Groper, 1996) Instead of informed debate, people made statements or pronouncements, often structured to humiliate others. (Benson, 1996, vanDijk, 1996)
A second theme in the literature is that there are characteristics of CMC that may help or hinder communication. CMC is not as rich in sensory information as face to face communication; therefore, in earlier research, online communication is seen as less socially oriented and more likely to be negative. Because people can participate anonymously, they can be more honest but they can also be more rude and attempt to dominate the conversation (Benson, 1996; James, Wotring, Forrest, 1995; Garramone, Harris and Anderson, 1986). Walther, Anderson, and Park (1994) reviewed this literature that found computer mediated communication to be cold and impersonal, and to encourage uninhibited communication (flaming); they described several shortcoming including the choice of groups to study, the artificial time limits, and the assumption that people communicate only face to face or online but not both.
Other research, however, looks beyond the workplace to the more social uses of communication technology and find different results. Rafaeli (1994) found evidence of playful behavior online and social rather than task oriented communication. Usenet groups bring together people who would not normally have the ability to meet and talk. (Baym, 1995) Anonymity may encourage people to open up and so connect users more directly. (Garramone, Harris, and Anderson, 1986) Participants create their own jargon, acronyms, and nicknames that add to the feeling of belonging to a cohesive group. (Baym, 1995) People use these unique vocabulary creations to give emotional content to their messages. The text itself does not convey the non communicative content of verbal messages - loud tones, significant pauses, arched eyebrows. (MacKinnon, 1996) With emoticons such as smiley faces, descriptive side passages, and attention to detail, text only messages can be enriched and much non-verbal content can be conveyed. There are nonverbal cues being used to enrich online communication and research needs to be aware of them (Walther, 1992).
A third them in the literature looks at why people participate in online discussions. Baym (1995) reviews some of the literature that points out that Usenet groups cover an incredibly wide variety of topics, most of which have nothing to do with professional or work related topics. People use the usenet groups to distribute information and to collaboratively interpret and extend that information as a way of enhancing their pleasure and enjoyment of whatever topic. Political CMC is thought to "promote better citizenship through more participation and communication" (Sakkas quoted in Hacker, Scott, Howl, and Steiner, 1996) because it reduces citizen reliance on traditional media, however one early study found that most of the people participating in online discussions were already the more politically active people. (Garramone, Harris, and Pizante, 1986) In 1985 people were using political bulletin boards for general political information and out of curiosity. (Garramone, Harris, and Anderson, 1986) Most political discussion is not work related; it is a voluntary leisure activity that many people find exciting. (Huckfeldt et al., 1996)
The users create the content in these newsgroups, set the agenda for discussions. (Benson, 1996) There are few if any outside editors or reporters. They post their opinions about the candidates; campaigns post information about events; and still other posters encouraged readers to take some action. (Hacker, Scott, Howl and Steiner, 1996) Larger communities may attract more lurkers but they also attract more posters because there is a better chance of finding someone with whom to communicate. Larger communities can also support a larger variety of purposes for being online and so attract more users. (James, Wotring, Forrest, 1995)
Methods and Data
For this project, several data sources were used. First, all of the messages posted, but not crossposted, from two usenet newsgroups, alt.politics.usa.republican and talk.politics.libertarian, from a different day of the week for a two month period were analyzed. Messages were selected from the month before and after a political event the researcher thought would engage the newsgroup participants were analyzed, using the search filter feature of the DejaNews search engine which allows the researcher to indicate a specific date and specific usenet newsgroup to read. Variables listed in Appendices 1, 2, and 3 were coded for each message. From the first group, alt.politics.usa. republican, messages were sampled from 10/1/96 to 11/26/96, which included 5 days before the 1996 presidential election and 3 days after the election. From the second group, talk.politics.libertarian, messages from sampled from 6/1/96 to 7/30/96, which included 5 days before the Libertarian Party national convention which chose their presidential candidate and 5 days including and after the convention. These groups were chosen to get a mix of partisan viewpoints; support for both major party candidates and several third party candidates, including Perot and Browne, were expressed by posters to both of these groups. There is little overlap in the people who were posting to the two groups.
To collect information on the participants, including demographic characteristics and rates of political participation in the real world, personal invitations that included the URL for the online survey were emailed to the authors of the sampled messages. Current participants of the two sampled newsgroups and other political newsgroups to which messages were crossposted were also invited, by means of a usenet post from the researcher, to take the online survey. Results from this survey are presented after the analysis of the usenet messages.
The usenet newsgroup community exists within 2 other populations: internet users in general and the general population. The data collected from these two usenet groups will be compared with these populations using the data from the Seventh Graphic, Visualization, and Usability Survey (GVU) of Internet users, collected online in April, 1997 (URL: http://www.gvu.gatech.edu/user_surveys/survey-1997-04/) and a subset of the 1996 National Election Study that includes respondents with internet access.
1996 Messages from alt.politics.usa.republican
520 messages were analyzed from this newsgroup. Although messages were posted all eight days, the number of messages posted per day dropped off after the election, with 74% (n=387) before the election and 26% (n=133) after the election. There were 115 new threads started in those messages. Two hundred people posted just once in the days sampled; they may have posted on other days. The most prolific person posted 32 messages. Thirty six people posted twice and 11 people posted three times.
The first characteristic considered was the general topic of the message: political, partisan, philosophical, other. 205 messages were partisan; they promoted a specific party or presidential candidate or voiced a negative opinion about a party presidential candidate. All of the messages could conceivably mention one or several of the candidates. Sixty three mentioned just President Clinton. Thirty nine mentioned just Dole. Ten mentioned just Harry Browne and two mentioned just Ross Perot. Other messages mentioned some combination of candidates: for example, 58 mentioned Clinton and Dole. 109 messages were political but not partisan. They discussed issues and political results of the issues instead of people. Only 10 messages were philosophical, usually discussing what liberalism or conservatism means.
The specific topic of the message was also coded. 172 of the posts dealt with some issue: political scandals (44), guns (15), abortion (13), online security (12), and taxes (11). Another large group of messages (156) dealt with some aspects of the campaign, such as performance during the debate or congressional campaigns. 48 news stories were posted. There were 47 personal attacks on other posters, including name calling, and suggestions to not post again.
Another variable for which the messages were analyzed was type of political activity in the message. More than 90% (n=471) of the messages were discussion (activity type = 3). Only 1.2% (n=6) of the messages could be considered spam (activity type = 8); this small number was not a concern among common posters.
There were some interesting although rare examples of other types of political activity besides discussion. One candidate issued an appeal on 11/2/96 for voters for his write-in campaign. He told what office he was running for and in what district and state, even including information about a major city in the district. Also on 11/2/96 one post advertised an organized political activity - a candidate's public appearance. The poster invited people to "come out and support the Republican Presidential candidate" A more positive action was suggested on 11/10/96. A poster provided the url for people with a grievance against the IRS to complain to a new government committee. Another poster suggested a not-so-positive political action on 11/26/96. "Want to have some fun? And support your president at the same time? Mail a dime to the Presidential Legal Expense Trust, Dept 70, Washington, DC 20055. And request a receipt."
Each message received a transaction code to describe the tone and type of communication; these are in Table 1. All of the interaction types were represented except 9 (asks for suggestion, n=0). Message type 5, gives opinion, was the most frequent (38%, n=198). These type 5 messages were 37% of the messages before the election and 39% of those afterwards; the election had no real effect on that type.
Table 1: Interaction Types by Date (percent by date) - alt.politics.usa.republican
| Date /Interaction | 10/1 | 10/9 | 10/17 | 10/25 | 11/2 | 11/10 | 11/18 | 11/26 | Row Totals |
| 1 | 2 2.2% | 1 1.2% | 2 2.3% | 7 9% | 5 10.6% | 1 4.2% | 18 3.5% | ||
| 2 | 6 6.6% | 1 1.2% | 1 1.2% | 3 3.8% | 2 2.9% | 3 7.3% | 2 8.3% | 18 3.5% | |
| 3 | 1 1.1% | 1 1.2% | 2 2.3% | 5 6.4% | 8 17% | 1 1.5% | 18 3.5% | ||
| 4 | 6 6.6% | 4 4.7% | 6 7.0% | 7 9% | 8 17% | 2 8.3% | 33 6.3% | ||
| 5 | 34 37.4% | 45 52.9% | 37 43% | 19 24.4% | 11 23.4% | 32 47.1% | 16 39% | 4 16.7% | 198 38.1% |
| 6 | 8 8.8% | 15 17.6% | 14 16.3% | 8 10.3% | 8 17% | 11 16.2% | 2 4.9% | 1 4.2% | 67 12.9% |
| 7 | 6 6.6% | 5 5.9% | 2 2.3% | 4 5.1% | 4 5.9% | 3 7.3% | 1 4.2% | 25 4.8% | |
| 8 | 8 9.4% | 8 9.3% | 2 2.6% | 1 2.1% | 2 4.9% | 21 4% | |||
| 9 | 0 | ||||||||
| 10 | 11 12.1% | 2 2.4% | 5 5.8% | 15 19.2% | 2 4.3% | 7 10.3% | 5 12.2% | 5 20.8% | 52 10% |
| 11 | 1 1.3% | 1 0.2% | |||||||
| 12 | 17 18.7% | 3 3.5% | 9 10.5% | 7 9% | 4 8.5% | 11 16.2% | 10 24.4% | 8 33.3% | 69 13.3% |
| Column Totals | 91 17.5% | 85 16.3% | 86 16.5% | 78 15% | 47 9% | 68 13.1% | 41 7.9% | 24 4.6% | 520 |
It is apparent that more people expressed disagreement (23.5%) than agreement (10.5%) with previous posts. Type 1 interactions (strongly agreeing with some previous post) were 3.5% of the total messages. One message that was coded as a type 1 began "Thank you for these posts. It's about time someone dealt with facts and reality." Overall, few of the posters expressed much support for one another. Type 12 interactions (strongly disagreeing with a previous post) made up 13% (n=69) of all the messages.People found more to disagree about after the election, such as the election results and interpretations of the "mandate" represented by the vote totals. The type 12 messages included personal attacks on other posters, such as "you socialistic anti-gun loons" or "you are just as deceitful as the clinton white house.." Some of the milder type 10 messages expressed their disagreement in logical terms that invited further discussion. For example, in response to someone who said they were going to vote for Perot, a poster said "Thereby giving Clinton a 'divide and conquer' your opponent's vote, and helping to possibly re-elect Clinton. With all due respect, NO THANK YOU!"
Messages were crossposted to a variety of groups, mostly other political groups. The most commonly crossposted to groups included alt.politics.clinton (65), alt.fan.rush-limbaugh (49), alt.politics.libertarian (45), alt.politics.democrat.d (44), alt.politics.misc (42), alt.conspiracy (37) and alt.current-events.clinton.whitewater (37), alt.impeach.clinton (36). Some of the nonpolitical groups include alt.flame (3), alt.recovery.addiction.sexual (1), alt.recovery.religion (1), alt.religion.christian (1), alt.evil (1), alt.women.supremacy (1), and rec.org.mensa (1).
1996 Messages from talk.politics.libertarian
This newsgroup got less traffic than does the alt.politics.usa.republican group. There were 145 messages posted during the sampling time period. 75 people (52%) posted only once on the days sampled. This is a larger percentage than on the alt.politics.usa.republican group (38%). The most prolific person posted 16 times which is 11% of the total messages. Many of the 61 new threads (42%) seemed designed as pronouncements rather than as conversation starters. There was no difference in the messages before and after the convention. In fact, only 5 of the messages dealt with the convention at all, and one of those was an antagonistic message posted by someone who seemed bent on disrupting conversation as much as possible.
It is also much less partisan than the republican newsgroup. More of the participants are philosophical libertarians (small l) rather than Libertarian party members (big L). Forty four messages were philosophical (30%), 33 were political (23%) and 26 were partisan (18%). 42 messages were non-political; this includes spam and personal attacks.
These messages were also coded using the Bales Transaction codes in Appendix 2. The results for this newsgroup are in Table 2. 45 of the messages (32%) gave the posters' opinions (activity type 5) and 37 provided some type of information (type 6). As with the republican newsgroup, there were more disagreeing posts (17%) than agreeing (10%).
Table 2: Interaction Types - talk.politics.libertarian
| Transaction Type | 1996 (N=145) |
| 1 - Shows Solidarity | 2 (1.4%) |
| 2 - Tension release | 7 (5%) |
| 3 - Agrees | 7 (5%) |
| 4 - Gives Suggestions | 10 (7%) |
| 5 - Gives Opinion | 46 (33%) |
| 6 - Gives Orientation | 37 (26%) |
| 7 - Asks for Orientation | 5 (4%) |
| 8 - Asks for Opinion | 2 (1.4%) |
| 9 - Asks for Suggestions | 1 (0.7%) |
| 10 - Disagrees | 1 (0.7%) |
| 11 - Shows tension | 4 (3%) |
| 12 - Shows Antagonism | 19 (13%) |
These messages were also coded for political activities, using the codes from Appendix 3. There were no political party activities - no fund raising for candidates (category 2), no announcements of rallies or public appearances (category 5). As with the alt.politics.usa.republican newsgroup, the bulk of the messages (83%, n=121) were category 3 - discussion. Also, as with the other newsgroup, one person posted the bulk of the outside news stories (category 4).
There were fewer crossposts than in the republican group. Approximately 41% of the messages were crossposted to at least one group. There was also a smaller set of groups to which messages were posted, most frequently tx.guns. As with the other newsgroup, though, most of the crossposts were to other political groups.
National Survey Data
The GVU surveys have been conducted every 6 months since April, 1994. Because there is no central registry of Internet users from which a random sample can be drawn, the GVU researchers advertise the survey as widely as possible. However widely they draw their sample, it is still a non-random probabilistic sample of people with enough Internet experience to find and complete the online survey.The average internet user was approximately 35 years old. 31% of Internet users were female. 45% were married. More than 50% have finished either a bachelor's or advanced degree. One third worked in computer related fields. The mean annual income was $58,000 and almost 50% of the respondents made over $50,000. Politically, approximately 41% say that they are Democrat and 34% Republican; the survey did not ask about third party affiliations. Almost 90% of the respondents are registered to vote and 60% of those actually voted in the last national election; 40% voted for Clinton and 25% for Dole. The younger users are much less likely to vote or to have a strong party identification. The GVU study did not ask about forms of political participation other than voting. (GVU, 1997)
The National Election Study is a national random sample survey conducted after each major national election. To better compare results with the larger population, the survey questions on political participation were taken directly from the 1996 National Election study. It seems likely that the group of NES respondents with Internet access (n=404)was the best group to compare with the usenet posters. They ranged in age from 18 to 79, with the largest percentage in their 30s. Thirteen percent have a high school diploma and 26% have graduated from college. Approximately 30% have some graduate credit. Approximately 85% of this group reported voting in 1996. The NES only asked about the major parties and Ross Perot; no questions asked about third parties.
Online Survey Results
Fifty five people responded to the online survey as of 26 February 1998. All but two were male; this seemed to be predominately true for the posters too during the time period of the survey, which makes the survey respondents different from the body of internet users as a whole as represented by the GVU survey and the NES Internet users. The respondents range in age from 16 to 61, though more people were in their twenties and thirties, which is similar to the respondents to both the NES and the GVU survey. 32% have household incomes below $50,000; 45% report incomes between $50,000 and $100,000; and 8% have incomes over $100,000, similar tot he results for the GVU respondents. This is a well educated group, better educated than the NES respondents with Internet access. 47% have some undergraduate credits or have an undergraduate degree. Another 16% have some graduate work and 31% of the respondents indicated that they have completed a graduate degree.
They have a wide variety of interests that they pursue online. The survey respondents post to more than 65 other newsgroups (34 Alt, 15 Rec, 7 Talk.Politics, 5 Sci, 2 Misc, 2 Comp, more than 20 others including community specific or specific newsgroup name not given) on a number of different topics. These were not just politically oriented groups, but also groups discussing historical food, alien visitors, dumpsters, paintball, the environment, and football groups.
They are active in their real world communities, again in a wide variety of groups, not just political ones. Only 2 people reported belonging to no groups.56% belong to one to four groups and 34% belong to 5 or 6 community groups. The NES Internet using sample is more active in their communities: 77% belonged to one to four groups. Survey respondents belong to a wide variety of groups in the community; for example, 67% belong to some issue related group and 51% belong to a sports group. Only one person reported belonging to a group for senior citizens, as we might expect given the young age of the respondents. This survey did not include questions about the depth of their involvement because it was designed to discover the breadth of usenet participants in other activities.
They are active politically as well; in fact, 42% were involved in 5 or more different activities in 1996. Voting and discussing politics were the most frequently cited political activities. 85% voted, 95% discussed politics with friends, and 78% tried to persuade a friend to vote for a specific candidate. In fact, 40% of the respondents discuss politics seven days a week. Only 3 respondents never discussed politics in their offline lives. Money is important to any campaign. 44% donated money to a candidate, 47% to a group that supported or opposed a candidate (such as a political action committee), and 35% to a political party. 38% went to at least one campaign rally and 36% reported displaying the candidate's name by wearing campaign buttons, displaying a yard sign or placing a bumper sticker on their car. However, the respondents are not as politically active as the Internet using NES sample group. Fifty one percent of the NES sample had done one other political activity besides voting and 88% had done at least 4 campaign activities.
Not all parties are represented in the survey respondents. More indicated that they felt closest to the Libertarian party. This was true on the republican list as well as the libertarian list. There were no supporters of the US Taxpayers party and only 1 supporter of the Greens party. There is some difference in the campaign activities that the party supporters undertook in 1996. All three Democratic respondents were very active; they participated in seven or more campaign related activities. The Republicans were less active; the only person to do no political activities was a Republican supporter. Activity in the community was dispersed over the different political parties. A Libertarian supporter was involved with 11 groups, but a Libertarian supporter was also involved in no groups in the community. The same pattern is true for the Republican supporters.
Discussion
This project was designed as an exploratory look at the participants of political usenet groups to see if they are merely collection of individuals or communities with characteristics similar to the traditional real world face to face community typically studied with social network techniques. If communities, then they would represent a new arena for investigation.
In general, posters used these fora to argue, criticize, and belittle. There were few discussions. Instead people made statements, issued pronouncements, and would not listen to any other interpretations of the ideas. As of 1996, they do not seem to be a tool for organizing or even describing political activity. There is a small group of people who consistently post. A few other people post occassionally. I saw no evidence of any moves to build a sense of belonging to a group, rather there were efforts to stop certain people for posting or calls to place certain email addresses into the "kill file" so their messages would be ignored. It also seemed that no one expected these fora to be a place to find campaign information. No one posted messages asking about platforms or candidate issue stands.
These two newsgroups do not exhibit many characteristics of a cohesive group, much less a community. The online world is unlike real communities in the sense that conversation is segmented into different topic organized online communities instead of occurring more spontaneously between people in a real world community. Online, usenet newsgroups are divided into topics and off-topic conversation is not typical. For instance, political messages do not show up very often in knitting groups. Posters have to make a conscious choice to go to a newsgroup that interests them. It is difficult to just overhear "conversations". The participants are argumentative, intent on making their point rather than engaging in informed discussion. Since users have to choose to go to the political newsgroups, it seems logical that these posters would be more strong in their attitudes and more extreme in their actions. These findings seem to reinforce those of Groper (1996).
Spam can represent a threat to the group as a community; people stop reading a group if it is too difficult to find substantive messages in a sea of get-rich-quick schemes and want-a-date personal advertisements. Both groups at the time saw few spam messages. During the time the newsgroup was being monitored for this project, it became clear how easily the "community" can disappear online. First, there was little overlap between the posters in 1996 and 1997. Many email messages are no longer active so posters can not be contacted. Second, in the middle of October, the newsgroup was inundated with "spam", though in this case it was not commercial messages, but rather off topic posts by a small group of people promoting National Alliance (Nazi) ideas. The volume of these messages increased from one or two a day to hundreds, including long tracts. Regular posters stopped responding to these messages and the libertarian content declined precipitously. After approximately a month, the interlopers moved on; there was no indication in the archived messages on why they had shown up or why they left. There also appears to be more commercial spam messages in 1997 after the National Alliance hi-jacking than before. These spam messages can also destroy community, as can be seen in the alt.sex newsgroups.
Crossposting can bring in ideas from a wider group of people; they are the closest equivalent to overheard conversations. It seems that crossposting could be thought of as the weak ties of the online world. (Granovetter, 1973) They allow readers to be aware of threads of interest to them in other groups that they might not normally read. This could strengthen the community by making it more informed and aware of outside possibilities and threats. However, in most cases, messages were crossposted to other political groups that already knew about the topic or that had firmly held but opposed ideas. The news stories that were cross posted might be the best example here of attempts to bring in new information.
Participants continue to be primarily male, well educated, and with higher income. These are the same user characteristics reported by James, Wotring, Forrest (1995) in their study of computerized bulletin board users. These characteristics are somewhat different from the larger group of internet users, which is, according to the GVU study, approximately 30% female. The GVU users are also better educated and have higher incomes than the general population. The participants are also active in their communities and in the 1996 campaigns. This indicates that using CMC at this stage is not bringing new people into political discussion, but rather giving already politically oriented people new channels for discussion.
Conclusions
This paper has described the type of communication that occurs on two political newsgroups and the people that participated in these discussions. It was not designed to be representative of all political newsgroups, and certainly not of all newsgroups. However, the findings suggest that at this point newsgroups are not actively used for political activities other than discussion. They seem to have had limited impact on political participation, except for information exchange and issue awareness. The attempts at more overt political action were met with silence - no follow up posts, including posts from the action instigator reporting any success. There were some interactions with bits of previous messages copied in for clarity; however, these were not conversations in the sense of several people exchanging a series of messages about a topic, as one might find in a more traditional social network. Rather there were concurrent monologues and dialogues, both of relatively short duration, perhaps a message or two long. They were more negative and argumentative than action oriented or informative. Several frequent posters seemed disconnected from the main theme of the newsgroup. This was especially true on the talk.politics.libertarian newsgroups where one user posted book chapters that no one ever referred to and another repeatedly posting the same argumentative and divisive questions. There were some attempts to have political philosophy discussions, but they quickly dissolved into arguments and name calling even on the more philosophical talk.politics.libertarian newsgroup.
Online surveying and email interviews might reduce the costs of dealing with a small scattered population. Using usenet gives the researcher access to a group of very politically active people, but also to people who hold fringe political views and are willing to express them, a group that is difficult to arrange access with in face to face communities because of their small numbers. However, the main focus of these people is discussion, or more accurately argument. At this time, these newsgroups may not be good places to research political action. Researchers should be careful not to generalize results from online studies to the general population or even to the general population of computer users.
Future research might follow message threads to get a different view of the interactions. This type of newsgroup, which is philosophical discussion oriented and argumentative, may not be the place to study the spread of political information through a community or the use of the Internet to spur political activity.
The Bales Interaction Categories seems to work well with online communication. One type of message that was difficult to code using this set of categories were those that used humor in a negative way to make fun of a poster. These messages were definitely expressing disagreement with the previous poster, but using humor, perhaps to defuse the situation. They tended to show up as some posters repeatedly posted what were seen as disruptive messages. The political activity categories from Rash were less successful in describing the messages in this particular newsgroup. The categories seem to presuppose positive behavior. Additional testing of these category schemes should be done on other newsgroups that are less discussion and more activist oriented. The results of this exploratory study though do seem to indicate that it will be possible to use existing transaction coding schemes with online communication.
Bibliography
Bales, Robert, Interaction Process Analysis: A Method for the Study of Small Groups, Cambridge, MA: Addison-Wesley Press, 1951.
Baym, Nancy, "The Emergence of Community in Computer -Mediated Communication" in Jones, Steven (ed.) CyberSociety: Computer-mediated communication and community, Thousand Oaks: Sage Publications, Inc., 1995, p. 138-163.
Benson, Thomas, "Rhetoric, Civility, and Community; Political Debate on Computer Bulletin Boards" in Communication Quarterly, v44 #3, Summer 1996, p. 359-378.
Bimber, Bruce, "Politics on the Net: Is there a Theoretical Foundation for the Speculation?" 8/2/96, online, URL: http://www.sscf.ucsb.edu/~survey`/net-politics.html
Bimber, Bruce, "The Internet and Political Communication in the 1996 Election Season: Research Note", 5/10/97, online, URL: http://www.sscf.ucsb.edu/~survey1/mobilize.htm
Garramone, Gina, Allen Harris, and Gary Pizante, "Predictors of Motivation to Use Computer Mediated Political Communication Systems" in Journal of Broadcasting & Electronic Media, v30 #4, Fall 1986, p. 445-457.
Garramone, Gina, Allen Harris, and Ronald Anderson, "Uses of Political Computer Bulletin Boards" in Journal of Broadcasting & Electronic Media, v30 #3, Summer 1986, p. 325-339.
Graphic, Visualization, and Usability's Seventh WWW User Survey, April 1997, online, URL: http://www.gvu.gatech.edu/user_surveys/survey-1997-04/
Granovetter, Mark, "The Strength of Weak Ties" in American Journal of Sociology, v78, 1973, p. 1360-1380.
Groper, Richard, "Political Leadership in Usenet Newsgroups", presented at the 1996 Annual Meeting of the International Society of Political Psychology"
Hacker, Kenneth, Max Scott, Lori Howl, and Robert Steiner, "Uses of Computer-Mediated Political Communication in the 1992 Presidential Campaign: A Content Analysis of the Bush, Clinton and Perot Computer Lists" in Communication Research Reports, v13 #2, p. 138-146.
Huckfeldt, Robert, Paul Beck, Russell Dalton, Jeffrey Levine, and William Morgan, "Ambiguity, Distorted Messages and Nested Environmental Effects on Political Communications", draft, 1996.
James, Michael, C. Edward Wotring, and Edward Forrest, "An Exploratory Study of the Perceived Benefits of Electronic Bulletin Board Use and their Impact on Other Communication Activities" in Journal of Broadcasting & Electronic Media, v39, 1995, p. 30-50.
Kling, Rob, "Synergies and Competition between Life in Cyberspace and Face-to-Face Communities", in Social Science Computer Review, v14 #1, Spring 1996, p. 50-54.
MacKinnon, Richard, "Searching for the Leviathan in Usenet" in Jones, Steven (ed.) CyberSociety: Computer-mediated communication and community, Thousand Oaks: Sage Publications, Inc., 1995, p. 112-137.
McLaughlin, Margaret, Kerry Osborne, and Christine Smith, "Standards of Conduct on Usenet" in Jones, Steven (ed.) CyberSociety: Computer-mediated communication and community, Thousand Oaks: Sage Publications, Inc., 1995, p. 90-111.
Parks, Malcolm and Kory Floyd, "Making Friends in Cyberspace" in Journal of Communication, v46 #1, Winter, 1996, p. 80-97.
Rafaeli, Sheizaf, "The Electronic Bulletin Board: A Computer-Driven Mass Medium" in Computers and the Social Science, v2, 1986, p. 123-136.
Rafaeli, Sheizaf and Robert LaRose, "Electronic Bulletin Boards and 'Public Goods' Explanations of Collaborative Mass Media", in Communication Research, v20 #2, April 1993, p. 277-297.
Rash, Wayne, Politics on the Nets: Wiring the Political Process, New York, NY: W.H. Freeman & Co., 1997.
VanDijk, Jan, "The Reality of Virtual Communities" in Groebel, Jo (ed.), Trends in Communication,1996, p. 39-63.
Walther, Joseph, "Interpersonal Effects in Computer-Mediated Interaction" in Communication Research, v19 #1, February 1992, p. 52-90.
Walther, Joseph, Jeffrey Anderson, and David Park, "Interpersonal Effects in Computer-Mediated Interaction: A Meta-Analysis of Social and Antisocial Communication" in Communication Research, v21 #4, August 1994, p. 460-487.
Wellman, Barry and Milena Gulia, "Netsurfers don't ride alone: Virtual Communities as Communities", 4/96, online, URL: http://www.acm.org/~ccp/references/wellman/wellman.html
Appendix 1: Variables Coded in Content Analysis of Newsgroup Messages
| Variable Name | Description | Coding Details |
| Counter | Sequential count | |
| Poster | Email address of person who posted the message | |
| Responding To | Email address(es) mentioned in the message | |
| Topic | Description of the topic of the poster's comments | |
| TopicCategory | Overall topic of the message | 1=Political, 2 = Partisan, 3 = Philosophical, 4 = Other |
| Tone | Overall tone of the message | 1 = Positive, 2 = Negative, 0 = Unclear |
| Quote | Whether text from another message is quoted | 1 = Yes, 2 = No |
| SigFile | Whether Poster has a sigfile | 1 = Yes, 2 = No |
| SigContents | Description of sigfile | |
| Candidate | Whether political candidate mentioned in the message or sig file | Blank if no, otherwise candidates names listed |
| Crosspost | Whether this message was crossposted to other newsgroups | 1 = Yes, 2 = No |
| Interaction | Code from Bales' Interaction Analysis coding scheme | 1-12, see Table 2 for details |
| ActivityType | Type of Political Activity | 1-7, see Table 3 for details |
Appendix 2: Robert Bales' Interaction Analysis Categories (Bales, 1951, p. 9)
| CATEGORY CODE | CATEGORY NAME | CATEGORY DESCRIPTION |
| 1 | Shows Solidarity | Raises other's status, gives help, reward |
| 2 | Shows Tension Release | Jokes, laughs, shows satisfaction |
| 3 | Agrees | Passive acceptance, understands, concurs, complies |
| 4 | Gives Suggestion | Direction, implying autonomy for others |
| 5 | Gives Opinion | Evaluation, analysis, expresses feeling, wish |
| 6 | Gives Orientation | Information, repeats, clarifies, confirms |
| 7 | Asks for Orientation | Information, repetition, confirmation |
| 8 | Asks for Opinion | Evaluation, analysis, expression of feeling |
| 9 | Asks for Suggestion | Direction, implying autonomy for others |
| 10 | Disagrees | Passive rejection, formality, withholds help |
| 11 | Shows tension | Asks for help, withdraws out of field |
| 12 | Shows Antagonism | Deflates other's status, defends or asserts self |
Appendix 3: Types of Political Activities that Could be Done Online (from Rash,Wayne Jr., Politics on the Net, 1997)
| Code | Activity Name | Description |
| 1 | OrganizeActivities | Ask for volunteers,ask for topics to be discussed, discuss upcoming activities, coordinate with other group's activities, try to attract new participants from groups with related interests and goals |
| 2 | FundRaising | Ask for money, donations, give address where to send donations, tell what donations will be used for |
| 3 | Discussion | General political discussion |
| 4 | News | Action alerts, breaking news stories, post press releases or pointers to news stories, plant false stories, spin events, provide media with details and story ideas |
| 5 | Advertising | Post dates and times of meetings and activities, calls for participation in upcoming conferences, talk about media appearances of candidates |
| 6 | ActionInstigation | Prompt others to act on action alerts, discuss how to do specific actions, encourage others to post results of actions |
| 7 | Communicate | With politicians who might be lurking on the newsgroup, to convey opinion or to urge their action |
| 8 | Spam | Commercial messages, off the topic of the newsgroup |
Appendix 4: Online Survey
This appendix has three parts - the message posted to the newsgroup and sent to posters to encourage participation, the study information sheet, and the actual survey.
Invitation Message:
I am a doctoral student at Indiana University. This semester for a class on social networks I am studying participation on politically oriented newsgroups. To collect demographic data and information on political activities of participants, I have created a survey which is online at php.indiana.edu/~kgregson/survey.html The results will describe to researchers used to dealing with face to face communities the types of discussion that takes place online and some characteristics of the people involved.
Please take a few minutes to complete the survey. If you have any questions, please email me at kgregson@indiana.edu
Thank you very much.