Research Report
Sense of Virtual Community in Listservs and Newsgroups
UNC
Department of Psychology
Introduction
The term “virtual community” can be used to describe any group of people who interact through computer communication (e.g., email, Usenet newsgroups). However, are they truly “communities” in the traditional meaning? One way to determine if virtual communities are like traditional communities is to examine their sense of community, their feelings of belonging, attachment and identity with each other. Sense of community has been called one of the essential components of a traditional community and researchers believe that virtual communities also experience a sense of community.
The purpose of this research is to determine if active listservs and newsgroups experience a sense of community. Particularly, it examines how feelings of community in virtual groups compare to traditional face-to-face communities. In addition, it examines what leads to the feelings of community that members experience. It is hypothesized that the following factors contribute to virtual groups’ sense of community:
· Participating in exchange of support in the group (i.e., asking and answering questions)
· Observing exchange of support in the group
· Knowing norms of behavior (i.e., what is appropriate behavior in the group)
· Participating in sanctioning (i.e., correcting inappropriate behavior)
· Observing sanctioning
· Creating an identity in the group
· Learning other member’s identities
· Trusting other members
· Believing that group members meet face-to-face
Additionally, it is hypothesized that members of different activity levels (e.g., lurkers vs. highly active members) have different senses of community and that different factors contribute to their sense of community. For example, active members may have a stronger relationship between participating in the exchange of support and sense of community whereas lurkers would have a stronger relationship between observing exchange of support and sense of community. Figure 1 graphically depicts the model tested in this research.
Figure 1: Model of
Sense of community
This report is divided into two main sections. The first section provides a detailed analysis of the groups who participated in this project. To protect the confidentiality of the groups, all names are pseudonyms and only the most general descriptions of the groups are used. The second section reports the results of testing the model. In this section, the groups’ data are combined into one dataset and analyzed.
Analysis of the Participating Groups
Group Information
Seven Internet groups were studied. Table 1 describes these groups, lists their social or professional orientation, and their computer mediated communication type (newsgroup or listserv). Additionally, it provides information about the number of participants who responded to the web survey from each group, gender distribution and age.
Table 1: Description
of the participating groups
|
Group |
Prof/Soc |
Computer type |
N |
Women |
Age |
|
Foodie List |
Social |
Listserv |
56 |
73% |
48 (10.2) |
|
Foodie News |
Social |
Newsgroup |
49 |
10% |
44 (11.1) |
|
Pet Fanciers List |
Social |
Listserv |
28 |
86% |
46 (12.1) |
|
Pet Fanciers News |
Social |
Newsgroup |
30 |
90% |
46 (13.1) |
|
Professional List |
Professional |
Listserv |
39 |
82% |
45 (9.8) |
|
Professional News |
Professional |
Newsgroup |
12 |
8% |
39 (10.1) |
|
Social Chat News |
Social |
Newsgroup |
37 |
76% |
59 (7.5) |
|
Unknown |
|
|
17 |
53% |
48 (13.6) |
|
Total |
|
|
268 |
63% |
48 (11.7) |
Note: The percentage of women shows the gender distribution of the groups. Because more women are present, their percentage is reported in the table. Age represents the mean (average) age with the standard deviation in parentheses. The “unknown” group represents 17 people for whom a virtual group could not be identified.
Two demographic differences emerged in this analysis: 1) the Foodie Newsgroup and the Professional Newsgroup had fewer women participantsthan the rest of the groups and 2) the Social Chat newsgroup was older than the other groups. Overall, the participants of this study were older and contained more women than is typical of much of the current Internet research.
Group Participants and Sense of Community
Table 2 shows how the different groups compare based on the types of participants within them and the groups’ sense of community. The types of participants were determined by the participants’ responses to how often they read their virtual group’s messages, how often they publicly post messages to the group and how often they privately email members of the group. Definitions of each group follow:
· Lurkers: Read messages but never or rarely post messages or email messages.
· Email Only: Read messages and email private messages to others, but do not post messages.
· Post Only: Read messages and post messages to the group, but do not privately email messages.
· Post & Email: Read messages, post messages and privately email messages to the group.
The sense of community measure comes from traditional measures of sense of community which have been modified to be appropriate for virtual communities.
Table 2: Participant
Distribution and Sense of Community
|
Group |
Lurkers |
Email Only |
Post Only |
Post & Email |
Sense of Community |
|
Food Lovers List |
21% |
2% |
29% |
48% |
3.13 (.38) |
|
Food Lovers News |
37% |
6% |
43% |
14% |
2.79 (.36) |
|
Pet Fanciers List |
32% |
7% |
25% |
36% |
3.08 (.43) |
|
Pet Fanciers News |
17% |
13% |
23% |
47% |
3.16 (.35) |
|
Professional List |
56% |
8% |
21% |
15% |
2.81 (.39) |
|
Professional News |
58% |
0% |
41% |
0% |
2.67 (.28) |
|
Social Chat News |
14% |
5% |
30% |
51% |
3.04 (.43) |
|
Unknown |
6% |
18% |
35% |
41% |
3.14 (.30) |
|
Total |
30% |
7% |
30% |
34% |
2.99 (.41) |
Note: All numbers represent the percentages, except for the sense of community which is the mean reported sense of community with the standard deviation in parentheses. Sense of community scores above 3 indicate groups that agree, on average, that they are virtual communities.
Several distinguishing features emerge from this table:
· The Professional groups have more lurkers than the other groups.
· Few members in any of the groups are exclusively email communicators.
· The Foodie Newsgroup and both Professional groups have very few people who both post and email messages.
· The Foodie Newsgroup and both Professional groups also have a lower[1] sense of community than the other groups.
Although it was expected that the professional groups
would have a lower sense of community than the social oriented groups, the
finding that the Foodie Newsgroup has a lower
sense of community than the other social groups is unexpected. It is worth noting that this group was the only social
virtual community in which few members participated by both publicly posting
messages and privately emailing to other members. This
group also had fewer women than the other groups. Because
of the nature of survey research, it is inappropriate to make causal claims
about the relationship between public/private communication or gender and
sense of community. Nonetheless, these findings suggest an area of research
to pursue in subsequent studies.
Testing the Hypothesized Model
The rest of this paper will discuss the hypotheses tests. To test the hypotheses, all of the groups’ information was combined into one large dataset and then divided by participant type (i.e., lurker, email only, post only, email and post). Then, a series of statistical analyses were run. Table 3 reports the sense of community for each of the different member types.
Table 3: Comparing
sense of community for the different participant types
|
Participant Type |
N |
Sense of Community |
|
Lurker |
78 |
2.65 (.29) |
|
Email Only |
18 |
3.05 (.33) |
|
Post Only |
81 |
2.96 (.35) |
|
Email & Post |
89 |
3.29 (.32) |
Note: N is the number in each group. The sense of community numbers represent the mean with the standard deviation reported in parentheses.
Statistical analyses show that Lurkers have a lower sense of community than the other participants. Also, the Email & Posters have a higher sense of community that the other participants. There is no reliable difference in the sense of community for the email only or the post only members. We can interpret these findings to mean that the least active members have a lower sense of community than all the other members and that the most active members have a higher sense of community than all the others.
Table 4 reports which of the hypothesized factors were found to actually contribute to sense of community. The purpose of this type of analysis was to determine for each of the separate member types which of the factors contribute to their sense of community. It was hypothesized that different factors would be important for the different types of members, and this was indeed the case. To interpret this table, examine each participant type’s row. An X under a variable means that that this variable contributes to the sense of community for that participant type.
Table 4: Hypothesis
Testing for each type of user and overall.
|
Type |
Observe Support |
Give Support |
Know |
Observe Sanctions |
Actually Sanction |
Create ID |
Learn ID |
Trust |
Meet FtF |
|
Lurker |
|
|
|
X |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
X |
|
Post |
|
|
XX |
|
|
|
|
XXX |
|
|
P&E |
XX |
|
XXX |
|
|
X |
|
XXX |
XX |
|
All |
|
XXX |
XXX |
|
XX |
|
|
XXX |
XX |
Note: X: weakly related; XX: moderately related; XXX: Strongly related. Blank means no relationship was found between this factor and sense of community for this participant type.
Interpretation of Hypothesis Tests
First, for the lurkers, the main factor which contributed to their sense of community is observing others sanctioning (i.e., correcting) other member’s incorrect behavior. Because the lurker’s sense of community is the lowest, this finding suggests that observing sanctioning may be a fundamental way to connect to a virtual community. Sanctioning is one of the most basic behaviors members do to maintain themselves as a “group.” This finding, therefore, may be evidence that an awareness of sanctioning within a group leads to the most elemental awareness of being in a community.
For the members who email only, it is their knowledge (or perception) that members of the group meet face-to-face which contributes to their sense of community[2]. Recall that there are very few participants who use only email as their primary way of communicating with others in the groups. It is therefore difficult to make a strong statement about this group and how their findings will generalize to members of this type in other groups.
Lurkers and email only members are the least “publicly” active of the member types; they read or email messages behind the scenes, but other group members do not know them. For the more public members (i.e., members who post messages or post and email messages), other factors become important.
In particular, for the public members, trust and knowing the norms of appropriate behavior are important. Trust may be important for these public members because they are putting themselves “out there” to the group. They have the potential to be “flamed” or otherwise ridiculed. Perhaps as their participation increases, their trust that the group will accept them increases. According to this analysis, as their trust increases, their sense of community also increases.
Additionally, these public members appear more likely to know and presumably follow norms of appropriate behavior. Like sanctioning, norms of behavior are one of the fundamental aspects of well functioning groups. Perhaps the difference between awareness of norms and awareness of sanctioning as it affects sense of community indicates a subtle distinction in a member’s attachment to the group. Lurkers realize that people are being called to task for behaving inappropriately whereas the active posters know what the appropriate behavior actually is.
The most active members are those who both post and email. This group has the highest level of sense of community and has the most factors contributing to their sense of community. In addition to trust and knowing appropriate behavior, these members also report a relationship between sense of community and a) observing the exchange of support within the group, b) creating an identity or “name” for themselves in the group, and c) the knowledge or perception that people in the group meet face-to-face.
There are several observations to be made about these findings. First, it is surprising that these members reported a relationship between observing the exchange of support and sense of community instead of participating in this exchange. These are clearly the people most likely to provide help. Perhaps for them, it is the idea that they are not the only ones helping which promotes a community feeling.
Second, it was expected that the most active members
would have a relationship between being known (“creating an identity”) and
feelings of being part of a community. More surprising
is the relationship between meeting face-to-face and sense of community. Because
the email only group also had this relationship, it is possible that private
emailing contributes to some sort of relationship outside the virtual group. This private relationship with others may then increase
their sense of community. Because of the larger number
of people in this most active group, it certainly helps bolster this interpretation
for the smaller, email only group.
The final analysis in Table 4 did not distinguish by
groups, but instead combined all of the participants together. This analysis strategy allows us to take advantage of a
wider range of variance among the hypothesized factors. This
analysis shows that there is a relationship between sense of community and
a) the amount of help provided (but not observed), b) knowing the norms of
behavior, and c) sanctioning (i.e., telling someone their behavior is wrong)
d) trust and e) meeting face-to-face. This analysis
demonstrates that the more people experiences of each of these variables,
the higher their sense of community regardless of their level of participation.
Practical Implications
This section provides preliminary implications for virtual groups based on the current research. Although sense of community is important for active virtual communities, it may not be essential. For example, the Foodie Newsgroup do not have a high sense of community, but they remain an active social newsgroup. Future research should more thoroughly examine the importance of sense of community on group functioning and member satisfaction. The implications presented here focus on groups who want to promote a sense of community among their members.
First, members need to have the ability to participate both publicly and privately to develop a sense of community. Public participation appears to increase the amount of trust among members. Private participation may increase feelings of more personal relationships between members although that assertion is not directly supported in this data.
Additionally, developers of and leaders in virtual groups should focus on developing and enforcing norms of appropriate behavior and providing support to members based on the purposes of the group. Although groups are not always aware of their norms of behavior, they are one of the keystones of effective groups, and according to this research, a potential starting point for sense of community. Both observing and providing support are also important parts of sense of community. Members participate in groups for particular reasons, often to share information and to learn. This exchange of support is important to develop a stronger sense of community.
Conclusion
This study examined active virtual groups to determine what contributes to sense of community for different types of group members. Not all the groups examined in this study should be considered virtual communities. Additionally, not all of the members within the virtual communities had strongly developed senses of community.
As may be expected, the most active type of members,
those who both post messages and email other members, have the highest sense
of community. The less active members are also engaged
in the virtual groups although to a lesser degree and for different reasons. Despite the differences in these members’ attachment to
and activity within the groups, they are all still members of the groups
and they likely play different but important roles. We must continue to seek
to understand the members’ self-imposed roles and their relationship to the
successful functioning of their virtual groups and virtual communities.