Figure 2-2
Figure 3-1
Figure 3-2
Figure 3-3
Figure 3-4
Figure 4-1
Figure 4-2
Figure 4-3
Figure 4-4
Figure 5-1
Figure 5-2
Figure 5-3
Figure 5-4
Figure 5-5
Figure 5-6
1. INTRODUCTION
How effective are chat rooms and message boards for group communication?
Most people are attracted to the Internet not as a place to read
news or to shop, but as a place to relate to other people who
have similar interests (Seltzer, 1995).
In May of 1996, I started a group called Stop Abuse For Everyone (SAFE). Stop Abuse For Everyone, as the name implies, is a group that aims to stop relationship violence and abuse of any sort. In particular, it concentrates on domestic violence against straight and gay men and lesbian women, because services for these groups have been so sparse.
After the group started, I set up a web page (at http://ursula.uoregon.edu/~jarubick/safe) that provides information about relationship violence and battered men. Initially, only a few visitors wandered onto the page, but after I registered the site with the major search engines, visitors started to pour in. A year later, more than 7,000 visitors have seen the SAFE web page.
The most intriguing part of the web page was the contacts and relationships that developed from it. I regularly correspond with reporters, psychologists, ministers, and students from all around the world about domestic violence, and I am a part of a network or community of people that are concerned with the subject. It was fascinating to watch these relationships form online, and, as a consequence, I developed an interest in group formation and communication on the Internet.
As Richard Seltzer states in the above quotation, what draws people to the Internet, for the most part, is to connect with other people. E-mail is probably the most popular use of the Internet, and is likely what convinces the majority of people to acquire Internet access. E-mail is, however, a person-to-person form of communication, and not designed for group interaction.
Programmers and designers have invented various ways of communicating as a group on the Internet: mailing lists, Usenet, IRC (Internet Relay Chat), chat rooms, MUDs (Multi User Dungeons), and MOOs (MUD Object Oriented). All of these are designed to enable group communication; however, little thought has been given to the effectiveness of these different mediums. The mentality behind the construction of each of these types of group models has been "if we build it and it is good, then people will use it." Little has been done to evaluate where and how each form of communication is effective, and to examine how groups are formed with these modes of communication on the Internet.
This thesis introduces the various forms of group communication
on the Internet, focusing in particular on a message board and
a MOO-style chat room, which are representative of the two major
styles of group communication presently available on the Internet:
synchronous and asynchronous.
Synchronous communication is analogous to conversation. A group
of people talk with each other in real-time, and respond to each
other's typing on the screen. Examples of this type of communication
are chat rooms, IRC (Internet Relay Chat), MUDs (Multi-User Dungeons),
and MOOs (MUD Object Oriented). Similar to conversation, synchronous
communication tends to be spontaneous, and a communication that
depends on responses and collaborative conversation.
Asynchronous communication is more similar to letter-writing
or note writing. The thoughts are composed, rather than spontaneous,
and the thinking is more structured. The intent of asynchronous
communication is often to convince, persuade, or inform. It often
involves a deliberate exchange of information. Examples of asynchronous
group communication are Usenet, e-mail mailing lists, and message
board systems.
The question this thesis will attempt to answer is "How effective
are chat rooms and message boards for group communication?"
Chat rooms are synchronous, and message boards are asynchronous.
In the course of this thesis, I will implement both of these
forms of communication, advertise their existence on the SAFE
web site, and compare the response the two forms of communication
elicit.
I will compare the two methods of group communication using the
following research questions:
1. Which form of group communication (chat rooms or message
boards) will people prefer to use? I will measure this by
the number of visits to the chat room and the message board, the
number of characters created in the chat room, and the number
of postings on the message board.
2. Which forms of group communication will draw people back
more often? I will measure this by the number of repeat visits
to the message board and the chat room.
3. With which form of group communication will people interact
more? I will measure this according to the number of conversations
on the chat room, and the number of replies to previous postings
on the message board.
4. Through which form of group communication will people become
better acquainted with other group participants? I will measure
this by means of a questionnaire to both the participants of the
message board and the chat room.
5. Which form of group communication will be more enjoyable
for the participants? I will measure this by a self-administered
questionnaire.
6. Which form of group communication will encourage group
formation more? This will be measured by all of the above
criteria. Specifically, the form of group communication that
has people interacting more often, has the participants know each
other, and draws people back more often will point towards a form
of group communication that encourages group formation. The literature
review gives an explanation of the importance of these factors
in group formation.
7. Will either form of group communication foster the growth
of an online community? Maltz, in Customary Law and Power
in Internet Communities, points out that a "sustainable
community requires continuity, and the sharing of public conversation
space" (undated). I will measure whether each builds online
communities by measuring the return rate of visitors to the message
board and chat room.
Different models of group communication on the Internet are valid
for different purposes. For example, MUDs offer an enjoyable
environment for group game-playing. However, I will be concentrating
on what is effective for a group interested in conversation, support,
and discussion about domestic violence.
Because this thesis aims to study two types of communities: a
chat-room-based synchronous community and a message-board-based
asynchronous community, it is useful to understand how groups
form, how they stay cohesive, and what factors influence their
development. Understanding these points will also help to design
the message board and chat room to minimize extraneous factors
that would hinder users from visiting the message board and chat
room. This understanding should help to make comparison between
the two forms of group communication more useful and accurate.
Thus, the first half of the literature review in Chapter 2 will
be drawn from social psychology, and discuss these questions.
The second half of the literature review will be a description
of the different forms of group communication on the Internet,
including the chat room and the message board, but also highlighting
the different variations. Because there are so many hybrids of
synchronous and asynchronous forms of group communication, this
will help show both the range and similarities within both synchronous
and asynchronous group communication.
Next, I will describe the chat room and the message board in
greater detail, followed by a description of the implementation
and data collection techniques used in this study. I will then
show the results of the study, and conclude with an analysis of
the data.
Because this thesis attempts to create two groups, a logical question
to ask is "What makes a person feel like they belong to a
group?" To answer this question, I will first review some
of the social psychology literature on group identity and membership.
Then, after looking at the difference between a group and a community,
I will examine the differences between face-to-face communication
and virtual/online communication, and see how that affects group
identity. I will conclude this section with a discussion of the
different forms of community on the Internet: Internet Relay Chat
(IRC), chat rooms, MUDs, MOOs, USENET, and online mailing lists.
Social psychologists have described group identity and its causes
in several ways. Bouas and Arrow (1996) sum up these views:
Existing perspectives on group identity can be grouped in three
general categories: (1) those that emphasize interdependence,
common fate, and collective interests (a behavioral or
experiential component); (2) those that emphasize cohesiveness
and the development of interpersonal bonds among group members
(an affective component); and (3) those that emphasize
awareness of the group and identification of oneself as a group
member (a cognitive component) (Bouas and Arrow, 1996,
p. 155).
The relative importance of each of these categories can vary depending
on the dynamics of the group, but looking at them all together
can be useful to understand why people identify themselves with
a group (Bouas and Arrow, 1996).
There is much more to group formation than group identity, however.
The question still remains of how groups and communities form
in the first place. McGrath defines small group formation as
"the emergence of a functionally interconnected whole from
a set of elements" (McGrath, Arrow, and Berdahl, forthcoming).
Groups may form for many different reasons, but the pressure or
force that initiates them can be divided into several different
kinds. McGrath identifies four different general forms of groups:
concocted groups, founded groups, emergent groups, and circumstantial
groups (McGrath et al, forthcoming).
Concocted groups are groups that are deliberately planned and
created. For example, a work group in an organization is a concocted
group. A founded group is formed when a several individuals link
up with other people who agree on a common purpose. An example
of this would be a neighborhood watch group. An emergent group
is a group that emerges spontaneously out of already existent
relationships. Friendships are often formed in this way. Finally,
groups that arise out of unexpected environmental situations are
called circumstantial groups. An example of this would be a group
of people stranded in a subway car together (McGrath et al, forthcoming).
The message board and chat room, if successful, will be examples
of founded groups, because the users are not forced into their
situation (as in a concocted group), the relationships are not
already present (as in an emergent group), and the relationship
is not haphazard (as in a circumstantial group). Knowing that
the SAFE groups will meet the definition of founded groups is
important, because they can then be compared directly with other
founded groups.
Several factors can make group formation more likely. Closeness
in a social network can be a good indicator of future ties among
people (McGrath et al, forthcoming). Unfortunately, the visitors
to the SAFE web page will not have this pre-existing social network.
Also, groups are more likely to form when "a change in the
environment opens up new opportunities, and at the same time,
either severs existing ties or reduces the demands of existing
groups for people and resources" (McGrath et al, forthcoming).
The SAFE groups could be said to be taking advantage of a new
opportunity, but the visitors are likely to not have severed any
previous ties or demands, unless they have just left an abusive
relationship. They actually may have an already existing set
of demands that will prevent them from participating in the new
SAFE groups. For example, an abused man may not feel like he
can read messages from the message board when he is at home and
his wife is awake.
The members of a group can also join for different reasons, and
this is important to note when forming these groups. The leaders
in the group may participate to fulfill a personal need for power
and status, or they and the other members may participate out
of an interest in a project the group will undertake, or a benefit
that being in the group can provide (emotional, social, or monetary,
for example) (McGrath et al, forthcoming). The goals of the group
can affect why people participate. For example, there is a large
difference between someone participating in a counseling group
and someone participating in a bomb squad. In the bomb squad,
the group formation is very task oriented, while in the counseling
group, the emphasis is on the interpersonal connections and the
relationship with the group (McGrath et al, forthcoming).
The chat room will probably focus more on developing interpersonal
connections than the message board. One reason this may be true
is that people introduce themselves in chat rooms, and talk about
themselves. On message boards, it is less common for people to
talk about their own personal experiences, especially from the
very beginning. People on message boards do not necessarily exclude
personal experiences from their conversations. I have seen personal
postings on message boards before. However, for the most part,
chat room conversation tends to be much more personal than message
board communication. This is a large difference between synchronous
and asynchronous communication.
Generally, people are more inclined to join groups with people
who share their own values and opinions (McGrath et al, forthcoming).
In the long run, diversity is beneficial for a group and allows
for more flexibility, but the cost of this is a longer start-up
period (McGrath et al, forthcoming). Based on my own subjective
observation of the SAFE web page visitors over the last year,
the visitors represent a diverse population of students, academics,
psychologists, abuse victims, and friends of abuse victims. This
will result in a longer start-up period, which, given the short
time period of this experiment, suggests the usefulness of a future
study that will observe group communication on the Internet for
a longer duration than the two months of this study.
All groups are not necessarily communities. Webster's defines
community as "a society of people having common rights and
privileges, or common interests, civil political, etc. or living
under the same laws and regulations" (Webster's, 1979).
The important parts of community are the social aspects and the
element of continuity which a group does not necessarily possess.
A study of The Well, an online message board system, defined
community as a "set of ongoing social relations bound together
by a common interest or shared circumstance" (Michalsky,
1995, p. 44).
The ultimate goal of the SAFE groups is to form two separate
communities. The most essential element of community is continuity,
and to measure whether or not the message board and chat rooms
have achieved any level of community, it will be necessary to
measure the continuity of interaction. This thesis will measure
how often people return to interact in the SAFE groups, on the
assumption that continuity in a group implies that the group is
a community. In the design of the chat room and message board,
I will also try to encourage people's continued participation
in the SAFE groups.
Online communication is such a new form of communication that
it does not offer a lot of comparisons to earlier forms of communication.
Because face-to-face group communication is familiar, comparing
face-to-face and virtual communication will help us understand
the strengths and weaknesses of this new medium, and help us design
our virtual groups as well as possible.
The primary difference between computer-based groups and in-person
groups is that computer-based groups tend to be less flexible
but more stable than in-person groups. If the environment does
not vary much, the structure of the group will be very solid and
stable, but computer-mediated groups will not deal well with sharp
or even large gradual changes. Arrow found that computer-mediated
groups with high membership fall into a very stable but rigid
structure if the membership did not change too much at the beginning.
Presumably, any sort of complication and fluctuation at the onset
could keep the computer-mediated groups from forming into a stable
structure (Arrow, 1997, p. 75).
A face-to-face group will survive changes in the group structure
that computer-based groups will not. Face-to-face groups are
slightly more flexible than computer-based groups:
Whatever structure emerges in a group's first interaction will
persist whether or not it is optimal, or even satisfactory, for
performance. Moderate changes in the environment will have no
effect on group structure. However when a shock to the system
jolts the group out of its usual fixed pattern, an abrupt and
radical restructuring will occur. The new structure that emerges
will then persist unchanged until the next crisis (Arrow, 1997,
p. 78).
As an example, consider a neighborhood watch group that gradually
adds members. The structure of this group will not change significantly,
because the change is gradual and "moderate". However,
if a new apartment complex was built in the neighborhood and the
membership doubled, then this could be a potential "jolt"
to the group, and the group may end up having "abrupt and
radical restructuring". If this same neighborhood watch
group were computer-based, the group would try to keep the original
structure, despite the large changes in group membership. Only
after a large breakdown would the computer-based neighborhood
watch group modify to fit its new environment. Face-to-face groups
will deal more gracefully with changes, such as group membership,
than computer-based groups (Arrow, 1997, p. 78).
The stability of computer-based communication likely arises from
the difficulty of establishing and changing the group's structure.
Computer-mediated group members identify less with their groups
than face-to-face groups (Bouas and Arrow, 1996, p. 153), and
the communication is often slower and inhibited by the difficulty
of learning the new medium.
One thing to note, however, is that all of this research on computer-based
groups was done on synchronous groups concocted by the experimenters.
Thus, the results of these studies might not have direct relevance
for the message board. However, it does suggest a pattern of
structure for computer-based and face-to-face groups that might
extend to asynchronous computer-based groups as well.
For the SAFE groups to acquire a stable structure, then, the
most critical time will be at the formation of the group. During
this time, Also, it will be important to strive for high membership
stability in the SAFE groups. If members shift in and out, or
the group's stability is shaken up for some reason, the group
will probably not establish a stable group structure.
Next, we will present the different models of group communication
and community currently available on the Internet.
Usenet is a vast forum on the Internet, composed of thousands
of different topics that people can discuss. The amount of information
going through Usenet probably qualifies it as one of the largest
exchanges of information on the planet. The amount of activity
on Usenet, and the thousands of newsgroups there, testify to the
amount of interest people have in Usenet. Maltz describes Usenet
in this way:
The most prominent of cybercommunities on the Internet is called
Usenet. Usenet is a hierarchy of conversation areas called "newsgroups."
In each of these newsgroups, users discuss topics by sending
a message to the group and reading responses. In many of these
newsgroups there is sufficient continuity and cohesion that each
can be considered a genuine and stable cybercommunity (Maltz
undated).
There are thousands of these newsgroups, filled with discussion
and information about various topics. For example, comp.lang.java
is a discussion group about the Java programming language, soc.culture.african.american
discusses issues affecting African-Americans, alt.drinks.jolt
is an area where people talk about Jolt Cola, and so on. Some
of these groups are inane, some scholarly, some informative, and
some social.
A person "posts" a message to a newsgroup, and this
message then is sent to news servers all over the world. These
news servers are computers that collect all of these posted messages
and display them when someone requests it. If a person browses
a newsgroup, they will see a "thread" in the conversations,
which means that messages that are about similar topics are linked
together. People who browse these places regularly will get to
know frequent posters. It is possible to "lurk" in
a newsgroup - always reading but never posting any messages.
Knowledge of the members of a group is important for group cohesion
(Arrow interview 1997, February 28), although it is often impossible
to know who is lurking in a newsgroup, readers can develop a sense
of group membership by knowing the frequent posters.
Usenet newsgroups are founded groups. For someone to form a
new newsgroup, they must find other people to support the creation
of the new group, and submit it to an open vote on the Internet
(an exception is the alt newsgroups). The types of conversation
can vary a lot depending on the newsgroup. Some newsgroups do
develop interpersonal communication (for example, the alt.support.*
groups may), and some newsgroups are formal and have no interaction
at all (for example, some newsgroups are strictly for information
dispersal and not for conversation at all).
Another area on the Internet where communication takes place among
groups is online mailing lists. The concept of a mailing list
(often called a Listserv or Majordomo list) is fairly simple.
Anyone with access to e-mail is able to participate in the conversation.
The person who starts a mailing list sets up a program on a computer
connected to the Internet. The computer maintains a list of people
that belong to the list. People who want to join the list can
write an e-mail to the computer in a specified format, and the
computer will automatically add them to the list. In a similar
fashion, the people on the mailing list can take themselves off
of this list. Any time a person on this list mails an e-mail
message to the central computer, the computer will reflect this
message to all of the people on the list. When someone replies
to your message, it too goes back to all of the people on the
list. In this way, a dialogue is possible, and conversation can
continue for extended periods of time, people can get to know
each other, and a group identity can even grow out of the communication.
A person who is on the mailing list is like a listener in a conversation.
They will receive the e-mail conversation and can read the conversation
as it progresses. If they choose to participate in the conversation,
then they can write their own e-mail message to the mailing list.
There are many variations to what I just described. There does
not actually have to be a central computer, and a list manager
can manually keep track of who is on the list and who is not.
Lists can be closed off so that people have to be invited or
have to be approved before they can join a list. Lists can be
moderated, so that all the messages have to go through a moderator
before they actually get posted to the mailing list. But generally,
a mailing list is an e-mail based conversation with a fixed or
slowly changing set of people, usually on a specific topic of
interest.
Usenet news groups and mailing lists are similar and actually
can even overlap. Some news groups are reflected on mailing lists,
so that members of the mailing list get copies of all the postings
to the news group. The reverse is also true - postings to the
mailing list are reflected onto the newsgroups.
MUDs (Multi User Dungeons) were originally developed as a form
of online gaming. They are a completely text-based adventure
game. In this game, each person that logs into the MUD assumes
a persona (for example, a female ogre warrior), with a set of
attributes (how the person looks, and how strong, quick, and intelligent
they are). The programmers set up a world of connected rooms,
which have various objects, obstacles, and monsters in them.
As the players move around in the rooms, collect treasure, and
kill monsters, they become more experienced and powerful as players.
They can cooperate and team up to kill a difficult monster, or
they can duel each other. There may be a form of currency. There
is also a rich form of communication. The players can talk with
each other, and nod, grin, laugh, grimace, hug, wince, etc…
at each other.
When a player or participant connects to a MUD, they give one-line
commands to the main computer (server) to tell their player what
to do. The computer parses and interprets the commands and informs
the player as changes happen in the environment. Thus, if a player
is in a room, they can "look" around the room and they
can "inspect" objects or other players in the room.
While they're in the room, other players or monsters may enter,
and people may engage in conversation or fighting. Many MUDs
also have elaborate spells, like "Teleport", "Heal",
and "Invisibility" which teleport the player, heal the
player, and make the player invisible to others, respectively.
MUDs are a form of role-playing, and often closely resemble games
like Dungeons and Dragons. There are a wide variety of genres
the game may take place in, and the focus of each MUD is different,
with some MUDs focusing on killing, and others on cooperation
and problem-solving. The overall focus, though, is having fun
with other people. Social contact is a central part of these
games.
MUD-based groups can be emergent and circumstantial. Some MUDs
have guilds or clans that players eventually can join. The clans
look out for each other, and share information, weapons, and magic
with each other. Often, the members of a clan will help out a
member that is in trouble. This type of group is emergent, because
the clans, rather than being founded, tend to evolve out of the
shared experience of being a player in the MUD. Relationships
can also be circumstantial, however. If a player happens to come
across another player that is losing a battle against a dragon,
for example, a makeshift alliance may be made that can last even
after the dragon is finished off. Players can team up to beat
a certain area of the MUD, or they can just team up for a brief
fight. For this reason, MUDs are both emergent and circumstantial.
They are also very task-oriented, for the most part.
MOOs (MUD Object-Oriented) are one of the offshoots of MUDs.
MOOs typically involve more conversation than MUDs, and their
purpose is usually more on conversation and communication than
tasks, such as killing and questing: "Its most common use,
however, is as a multi-participant, low-bandwidth virtual reality"
(Curtis, 1996, p. 1).
One aspect of MOOs that differentiates them from MUDs is the
presence of complex objects. For example, it is possible to create
a paper object that can be folded into a paper airplane and thrown
in a room. The airplane, according to how it is programmed, could
then fly around the room for a few minutes and then crash into
a wall and end up as a broken paper airplane. The objects can
often be quite sophisticated, and this makes a MOO a rich environment.
With enough programming, the participants can interact with their
environment a great deal.
Participants in some MOOs have developed online communities of
sorts. Dibbell (1993) discusses how several players were "raped"
online on LambdaMOO, the MOO operated by Xerox Parc (located at
telnet:lambda.parc.xerox.com 8888). The victims of this
assault felt extremely violated, and the rest of the LambdaMOO
participants felt outraged. They discussed at great length what
to do about this and how to prevent future assaults. Dibbell
was one of the observers in the MOO at the time, and his comment
was that this was truly a community. The members of the MOO was
actually setting up a system of self-governance, almost a simple
type of government. McGrath would probably define this type of
group as a founded group.
In some ways, IRC (Internet Relay Chat) and chat rooms, two similar
technologies, are similar to real-time versions of Usenet newsgroups.
As on Usenet, there are groups that you join to participate in
a conversation. Groups in IRC are called channels, and rooms
in chat rooms.
The primary difference is that the conversation is not a composed
message which other people respond to. On IRC, the unit of conversation
is a line of text, as shown in Figure 2-1.
/join #pix
*** irc-2 (~jarubick@ursula.uoregon.edu) has joined
channel #pix
/who #pix
#pix kamera H ~jarubick@ursula.uoregon.edu
(Jade)
#pix Mem H ~Mmr@host1.msp.com (figure
it out)
#pix ur1 H kali@ind-0007-19.id.net
(Katherine R)
#pix ride H@ jd@fn-ppp47.ls.net (tony
b)
<ride> kamera age sex
> 24 M
<Mem> 15 female
<ride> what kind of pics wanted
<ride> ur1 getting it
> Sorry just messing around
<ride> its agood 1
<Mem> hey
Figure 2-1Figure 2-1
In this figure, I joined a group, looked to see who
was in the group, and then started participating in the conversation.
A person named <ride> asked me my age and sex immediately
after I logged in, and I replied 24 M.
Thus, the people in a room can type in a line of text at a time,
and read what everyone else has written. This allows for a conversation
that can proceed fairly naturally, provided the participants can
enter the text into the computer with enough ease to allow the
flow of conversation to continue. IRC can be fairly arcane and
complicated for the uninitiated. If you are not using a client
program (like Palace or Homer), the interface can be daunting.
Figure 2-2 illustrates what the user is presented with when logging
into an IRC session.
The immediacy of IRC is its greatest appeal. A channel can have
a small or large number of participants. In the larger groups,
and on many of the online service providers' (such as American
Online and Compuserve) chat rooms, the discussion often seems
disjointed. Several conversations can be happening at the same
time, and it often appears to resemble a room at a party or social
event, with smaller groups of people clustering off to have their
own conversations.
Chat rooms have been very successful for companies such as America
Online, and are very popular. People can consume great amounts
of time on IRC and in chat rooms, and they seem to enjoy socializing
online, meeting and talking with people they have come to know
through their online interactions. I know people who have started
relationships with others that they met on IRC, and I know one
couple that married after becoming acquainted on IRC.
For the most part, chat room participants are looking for communication
and interrelatedness, rather than for solving a particular task.
Also, the form of group is usually emergent.
The two types of group communication, synchronous and asynchronous,
are completely different types of communication. In order to
test which form of Internet communication is more effective for
communication and group formation, I set up one of each: a MOO
style chat room and a message board. The following section describes
each in great detail.
The chat room was built around a MOO (MUD Object Oriented) developed
at Xerox Parc, called LambdaMOO. This MOO is written in C and
is still operational at the time this thesis was written. It
is accessed from the SAFE home page (http://ursula.uoregon.edu/~jarubick/safe).
The MOO consists of the code for the MOO (the server) and a database
taken from the original LambdaMOO. The database contains information
on all of the objects in the virtual reality, including all of
the programming for the objects and their descriptions, as well
as help information, basic commands, and information to help create
the environment for MOO programmers. The database starts out
at 2143Kb, and the MOO program, or executable, is 919Kb. The
database will gradually grow as more players, objects, and rooms
are added to the MOO.
Adding rooms and objects is the job of the "Wizard",
or MOO programmer, and requires a significant time investment.
The MOO has its own programming language that Wizards can use
inside the MOO environment. The object programs are object-oriented,
and can inherit the properties of already existing objects. For
instance, as the programmer, it is possible to create a "schedule
book" that is based on an already existing "notebook"
object. The schedule book would have most of the properties of
the notebook. Thus I could "write" in the notebook
and "erase" entries in the notebook, but at the same
time, I could change or add the properties of the schedule book
- for example, maybe the schedule book could remind me of an appointment
an hour beforehand.
As might be deduced, verbs are associated with objects. If a
programmer creates a new object, the programmer must also create
most of the verbs that participants can do on that object. If
a programmer makes a "cup of coffee", the programmer
also needs to make sure to program in what will happen when the
player "drinks" from the cup of coffee. Perhaps the
"drink" program will change the name of the cup of coffee
to an "empty cup". In this case, the act of drinking
from a cup of coffee will result in an empty cup.
Once the program is set up, it is accessible from any computer
that is connected to the Internet and has a Telnet program. The
address for SafeHouse is zfishstix.cs.uoregon.edu 7777,
meaning port 7777 of Zfishstix. The program runs constantly as
a server program, and all the programs that connect to it act
as clients.
The message board is accessed through the SAFE home page (http://ursula.uoregon.edu/~jarubick/safe)
using any web browser, although posting requires a browser that
supports forms (as almost all browsers do) such as Netscape Navigator
2.0 or later. The main web page is accessed from the Stop Abuse
For Everyone home page. As Figure 3-1 shows, directions for the
use of the message board are shown at the top of the page, readily
accessible to any newcomers, and previous messages are shown at
the bottom of the screen. Any message that is a reply to another
message is indented underneath the original message. The date
that the message was posted is displayed after the title of the
posting.
At this point, the user can click on any of the messages to read
them, or click on the Post Message button on the bottom of the
screen. If the user clicks on the Post Message button, then they
are presented with same screen as Figure 3-2. The name and e-mail
address sections are both completely optional, as is explained
in the directions page. The Check Here option allows users to
have e-mail messages sent to them when another person responds
to their messages. The buttons at the bottom of the screen, Post
Message and Clear Message, are for the user to post the message
they compose, or clear what they have written and start over.
If the user instead decides to read a message and clicks on one
of the messages, the message will show up, similar to Figure 3-3.
The user can then read the message and, if they choose, reply
to the message (Figure 3-4). If the user would like to privately
respond by e-mail, they can click on the e-mail address (providing
that an e-mail address was given by the poster of the message).
Notice that links can be embedded inside of a message to other
places on the Internet. In this example, on Figure 3-3, I wrote
a link that users can click on to go to the main Stop Abuse For
Everyone (SAFE) homepage.
This interface was developed with the assistance of the Human
Computer Interaction group in the Computer Science Department
at the University of Oregon. I adopted as many of their suggestions
as I could, with the goal in mind of reducing any factor which
could distract the visitor from easily using the message board.
Figure 3-1Figure 3-1
Figure 3-2Figure 3-2
Figure 3-3Figure 3-3
Figure 3-4Figure 3-4
In this empirical study, I studied two types of Internet group
communication to observe the development of group communication
and formation: synchronous and asynchronous. A message board
and a MOO-style chat room were connected to the Stop Abuse For
Everyone (SAFE) website, and I studied the effectiveness of these
two group communication models.
Specifically, I looked at the following questions:
1. Which form of group communication (chat rooms or message
boards) will people prefer to use?
2. Which forms of group communication will draw people back
more often?
3. With which form of group communication will people interact
more?
4. Through which form of group communication will people become
better acquainted with other group participants?
5. Which form of group communication will be more enjoyable
for the participants?
6. Which form of group communication will encourage group
formation more?
7. Will either form of group communication foster the growth
of online community?
In the following section, I will explain the process I followed
to set up the message board and chat room, and show the questionnaires
and data collection methods I used to answer these research questions.
I created the chat room using a MOO obtained from Xerox Parc's
FTP (file transfer protocol) site (ftp://parcftp.xerox.com/pub/MOO/).
The software is public domain. There are several different versions
of the software and the database file used by the program, and
I downloaded version 1.8.0p5 of the MOO and the 02Feb97 version
of the database.
I set up the chat room on the Human Computer Interaction group's
computer, Zfishstix. To set up the program, I
1) ran the configuration program "configure"
2) edited the config.h file to make sure the options were set
up correctly for Zfishstix and the University of Oregon's network
system.
3) compiled the MOO.
At first, the program would not compile correctly, but after
reloading the source code from the FTP site, the installation
worked smoothly.
After the program was set up, the next step was to design the
internals of the chat room - the environment that the visitors
would interact with. This was a very important decision. Many
of the visitors would presumably be people with first-hand experiences
with domestic violence. The difficulty in choosing an environment
or theme for the chat room was compounded by the diversity and
range of possible environments. The chat rooms could be a small
town in a forest, several rooms in a high-rise building, or a
house. It could even be completely fantasy-like - a castle on
top of pink clouds, for example. The environment could have a
large impact on what type of experience visitors would have and
could impact whether they would feel comfortable interacting with
other people there. The environment was also important because
part of the purpose of setting up the chat room is to decide on
what is effective and what is not effective about chat rooms in
general. If the environment hinders people from using the chat
room, then I would not learn anything about the effectiveness
of the chat rooms themselves, other than the fact that the environment
hindered people from using it.
First of all, I conducted informal interviews with about ten
people, including architecture students, psychology majors, and
user interface specialists from the Computer Science department.
The questions I asked them were, "What type of an environment
do you feel most comfortable talking to strangers in?" and
"What makes a person feel comfortable in their environment?"
After receiving a lot of recommendations on how to design a comfortable
environment, I decided on a beach cabin as the theme for the chat
room.
The beach cabin was designed to be cozy and comfortable - the
type of place that people would want to talk in. It is dimly
lit, but not dark. The coffee machine in the kitchen is set there
because people tend to talk more over food or drink. The chairs
are comfortable and present in every room of the SafeHouse. If
the visitors feel claustrophobic, they can go out to the beach.
The basic room descriptions are listed in Figure 4-1, and the
tutorial rooms are listed in Appendix A.
Room Descriptions:
Entrance Way
You are in the entrance way to a cozy beach cabin. Behind you,
to the south is the sandy beach, and to your north, you see a
warmly lit living room. If this is your first time, type tutorial.
Kitchen
The kitchen holds a dark wooden table with chairs clustered around
it. You can go back to the living room through the hallway to
the east. (the coffee maker is kept in the kitchen, usually,
and would be displayed as well)
Living Room
The living room centers around a large fireplace. The room seems
crowded with comfortable chairs and two large couches. The room
is lit by some warm lamps in the corner and the fire in the fireplace.
You see a hall leading to the kitchen on the west, a den to the
north, and the entrance way on the south.
Den
This actually looks like a library, except it is so comfortable.
The dim lighting hides the titles of all the books lining the
walls. A desk sits in front of a small window. Several plush
cozy-chairs are circled around. This looks like a great place
to relax. The living room lies to the south.
Beach
The surf pounds the sandy beach. This beach is an alcoved area,
with cliffs rising on both sides. This area seems sheltered,
but a light breeze runs in from the ocean. The salty air carries
seagulls through the sky. You can go north back into the cabin.
Figure 4-1Figure 4-1
I also pared down the commands as much as possible to make it
easy for the user to acclimate himself to the environment. The
default introductory screens and information screen (that is shown
when the user types news, as prompted when they first log in)
were both fairly technical and confusing. I pared down the commands
to just a subset of the commands available, because the full functionality
of the MOO is not necessary for someone who just wants to talk.
I then edited these introductory messages to make them more understandable
to people who have never used a MOO before.
Then, I set up a web page to link in from the Stop Abuse For
Everyone homepage to the SafeHouse chat rooms. Figure 4-2 illustrates
text of the webpage (http://ursula.uoregon.edu/~jarubick/ safe/MOO.html).
SafeHouse: Chat Rooms
You've probably heard people talk about chat rooms
before, and perhaps you've tried them yourself. SafeHouse is
a house of chat rooms - several chat rooms linked together. You
can use SafeHouse as a place to actually meet and share experiences
with other people whose lives have been touched by domestic violence.
Whether you are the person who is abused, or you know someone
who is in a similar situation, SafeHouse is an area for you to
talk with other people and learn from their experiences.
The SafeHouse chat rooms are available because I
am doing my thesis on group communication on the Internet, and
I'm comparing chat rooms and message boards. So please complete
a SafeHouse Survey (http://www.cs.uoregon.edu/~jrubick/BBS/MOO_questionnaire.html)
after you have tried it out. It takes less than five minutes,
is completely anonymous, and will help researchers understand
and design better methods of communication in the future. If
you have any questions or concerns about this, feel free to contact
me (jrubick@cs.uoregon.edu) or my thesis advisor (douglas@cs.uoregon.edu).
To use SafeHouse, you must have a "telnet"
program on your computer. If you'd like to try to connect to
SafeHouse, select the link below. If you have telnet, and your
web browser (like Netscape) is set up right, clicking on the link
below will allow you to connect to the SafeHouse.
Enter the SafeHouse: telnet zfishstix.cs.uoregon.edu
7777
The SafeHouse will not be very interesting unless
other people are on. Jade will be on almost every evening around
9:00 Pacific time. If you go into the SafeHouse, you can leave
a note for everyone else on the Message Board, and say when you
plan to connect again. People can make appointments in this way.
Please note: please be courteous and respectful towards
the people you interact with on SafeHouse. The people you are
talking with are real people. Please help to make SafeHouse as
comfortable a place as it can be.
If you have any suggestions, concerns, or complaints,
please either talk with me (Jade or Wizard) on SafeHouse, or e-mail
me at jrubick@cs.uoregon.edu.
Figure 4-2Figure 4-2
A CGI (Common Gateway Interface) script is a program that can
be run when someone accesses a web page. For example, if I write
a CGI program that displays the time, then when a person accesses
a web page with that CGI program, they will have the current time
printed on their screen. CGI programs are very common on the
World Wide Web, and they are used for all sorts of purposes, Message
Boards being one of them.
Selena Sol maintains a website that contains a great number of
public-domain CGI scripts. This website, currently located at
http://selena.mcp.com but soon to move to http://www.extropia.com,
is where I acquired the code for the message board system used
on the Stop Abuse For Everyone (SAFE) page.
The scripts themselves are written in Perl, a computer language.
To set up the scripts, the server must not only have Perl installed,
but the protection levels of the HTTP (Hyper Text Transport Protocol)
daemon's (a small program that runs constantly on a computer)
security level must be set so that it allows programs to read
from other programs while executing. This is necessary because
Selena's Message Board reads from other source files in order
to execute, and on some systems where security is very important,
this is not allowed by the HTTP daemon.
After downloading the program, following the setup instructions
resulted in the program working fairly smoothly. The most complicated
part of the process was setting up the file permissions. It took
several days to figure out exactly how to set them up, because
the documentation for the message board is not very specific or
helpful about the installation process.
After the program was working, I modified the code to customize
the appearance and display of the message board. The Human Computer
Interaction group, specialists in user interfaces, also informally
examined the message board and made many suggestions to make the
interface as unobtrusive as possible for the user. I implemented
as many of the suggestions as possible, and the end result is
shown in Figure 3-1.
1. Which form of group communication (chat rooms or message
boards) will people prefer to use?
Measure: number of visits
The log file from the MOO shows when each participant logs into
the chat room and when they log out. This is helpful because
it shows how often people logged into the MOO, how often they
came back, and how long they stayed in the chat room before leaving.
We can compare the number of visits to the chat room to the number
of visits to the message board, and thereby judge which form of
group communication people prefer to use.
The message board also logs all of the visits to each page.
As a result, the log files from the chat room and the message
board can be easily compared to see which is used more often.
On a secondary level, we will be able to judge the effectiveness
of the chat room based on the number of characters created in
the chat room and whether or not the visitors has described themselves
in the chat room. If the person has described themselves, then
they have demonstrated a mastery of the basic commands of the
chat room, and they have spent some time following the tutorial
and learning the commands.
2. Which forms of group communication will draw people back
more?
Measure: repeat visits
The log files will show the number of repeat visits, so I will
use this information to evaluate which form of group communication
draws people back more.
3. With which form of group communication will people interact
more?
Measure: number of conversations (chat room), number of
replies (message board)
In the chat room, overlapping log-in times in the log file represent
times when more than one person is in the chat room. Because
SafeHouse is so small, I will assume that if more than one person
is in SafeHouse, they are conducting a conversation. Using the
log files from the chat room, I will count the number of conversations.
On the message board, I will count the number of replies to previous
messages, which represent an interaction.
4. Through which form of group communication will people become
better acquainted with other group participants?
Measure: Questionnaire, question 5.
I will measure this by means of a questionnaire to both the participants
of the message board and the chat room (see Figure 4-3 and 4-4).
5. Which form of group communication will be more enjoyable
for the participants?
Measure: Questionnaire, questions 7 and 7b.
I will measure this by a self-administered questionnaire (see
Figure 4-3 and 4-4).
6. Which form of group communication will encourage group
formation more?
Measure: Level of interaction (number of conversations
for the chat room, number of replies for the message board), knowledge
of other group members (questionnaire, question 5), and drawing
people back (repeat visits to the web site or to the MOO).
7. Will either form of group communication foster the growth
of online community?
Measure: repeat visits
The log files will show the number of repeat visits, so I will
use this information to evaluate which form of group communication
draws people back more.
I placed two questionnaires on the web page for people to fill
out, one for the message board and one for the chat room. A sample
question is shown in Figure 4-3. As can be seen, the user can
select from multiple choices using the standard web browser interface.
Figure 4-3Figure 4-3
The full text of questionnaire is shown in Figure 4-4, and the
entire questionnaires, with all the choices available, are shown
in Appendix B (the actual page is available at http://www.cs.uoregon.edu/~jrubick/BBS/
BBS_questionnaire.html).
1) Gender:
2) Age:
3) How many visits have you made to the SafeHouse?
4) How often do you visit this site?
4b) If you wrote other in the last question, please indicate why
5) Approximately how many people do you know in the SafeHouse?
6) Do you feel like you're participating in a group:
6b) If you have any comments about the previous question, please
write them here:
7) How was your experience participating in the message board:
7b) If you have any comments about the previous question, please
write them here:
8) Is this your first time to fill out this questionnaire?
9) Please write any addition comments about this questionnaire
or the SafeHouse:
Figure 4-4Figure 4-4
The questionnaires for the chat room and the message board are
almost identical, but there are a few differences to reflect the
differences in the mediums. Both of the questionnaires are shown
in full in Appendix B.
During the two months of observation, the cumulative number of
visits to the chat room was fifteen, or twenty-eight including
partial log-ins, where the user does not connect to a player.
The reason the partial log-in figure is important is that a person
can connect to the chat room and check whether anyone else is
present without connecting to their player (a "full log-in").
Thus, the twenty-eight figure is probably fairly accurate. When
this is contrasted to the 201 cumulative visits to the main message
board page (see Figure 6-1), it is clear that many more visits
were made to the message board than the chat room. During the
two months of observation, there were 1503 visits to the main
web page, so this means that about one in 7.5 visitors went to
the message board and about one in 30 visitors went to the chat
room (because the chat room started up a week and a half later
than the message board did.)
There were only nine characters created in the chat room, including
one pre-created guest character. Of these nine characters, only
one logged in more than once. Three of the eight characters progressed
far enough through the tutorial and the learning of the chat room
to describe their own characters. These people likely either
lost interest or could not understand the technology and environment.
There were thirty-two postings for the message board (see Figure
5-2).
Figure 5-2Figure 5-2
There were many more visits to and use of the message board system
than to the chat room. This suggests that users preferred to
use the message board much more than the chat room. The postings
and character creation statistics are not conclusive or meant
to be compared to each other, but they serve to show further that
the chat room was not used as much as the message board.
There are a number of possible reasons for this. First, one
major difference between the two implementations is that the chat
room needs to be accessed by a Telnet program, while the message
board system is accessed by a web browser. Two of the questionnaire
respondents mentioned this as a reason they did not use the chat
room. Because the visitors to the SAFE website are visiting by
means of a web browser, it may be easier to look at the message
board than to visit the chat room.
However, studies have shown that users of any Internet based
service (such as the World Wide Web) use most of the other types
of Internet based services as well (Kraut, et al, 1996, p. 286).
This study was based on residential Internet use, though, and
it is difficult to tell what percentage of the visitors to the
SAFE groups were residential. Also, most browsers allow a user
to click on a "Telnet link" to automatically load up
Telnet when the user selects a link to a Telnet site. I added
in this type of a link to the page that describes the chat room,
so users should be able to easily enter the chat room (See Figure
5-3).
Figure 5-3Figure 5-3
Another possible reason for people feeling more comfortable on
the message board than the chat room is the nature of the participants.
First of all, the participants may have felt more comfortable
in the message board environment because the interaction is more
distant and safe (in the same way that writing a letter is different
than talking on the phone). This may be especially valid considering
that many of the participants are abuse victims, and concerned
with their privacy and anonymity. Although both mediums of communication
were publicized as anonymous, the message board may feel more
anonymous than the chat room, because the user is not immediately
talking with anyone. For people setting up web site services,
this suggests that the type of users may be a very important factor
in whether or not a service (such as a chat room or message board)
is used or not. Perhaps the chat room would have been more successful
if the participants were conversational and looking for human
connections.
Perhaps in addition to the feeling of safety, the motivation
for visiting the SAFE homepage is of most importance. If the
participants were motivated to come to the web page by a desire
to find information, they may not be looking for personal connections
with other people in similar situations. Visitors to the SAFE
web page chose to look at the page on the chat room seventy-nine
times (see Figure 5-4) versus 201 times (see Figure 5-1) for the
message board. Of the seventy-nine visits to the web page about
the chat room, only nine people chose to participate. Because
visitors would not know they needed a Telnet program until they
looked at the web page about the chat room, this suggests that
the visitors to the SAFE web site were simply not as interested
in using a chat room as the message board.
Figure 5-4Figure 5-4
Also, The message board is more suited for information exchange,
and this may have biased the visitors towards using the message
board. The message board works better for information exchange
because it is asynchronous. As a result, the information is permanent
(the visitor can visit the web page at a future time and get the
same information again, while with a chat room, the information
must be written down before it disappears, because chat rooms
do not store any history), and also the information can be exchanged
asynchronously, so it is not necessary to find someone who is
using the computer at the exact same time.
Thus, studying the composition and motivation of visitors to
a web site may be an important step before implementing any sort
of group communication technology on the Internet. In the same
way that a market study is important for a business, perhaps a
user study and an analysis of the people who will be using the
group communication is critical for the success of the group communication
method.
Users chose not to return to the chat room, and came back to the
message board much more often: approximately forty-three of 167
visits, or 25.7%, were return visits for the message board. If
I subtract out one user who returned 23 times, the percentage
drops to 14.0%. The chat room had only one person out of eight
(12.5%) return even for a second time, so if I exclude this person,
I arrive at a figure of 0% to compare with the 14.0% of the message
board.
Note that there are some difficulties in analyzing the message
board log data. Proxy servers, servers that create a different
Internet address for a user each time they connect to the Internet,
make it difficult to tell if a user is returning for another visit
or if they are actually a different person. For example, America
Online, a popular online service, uses a proxy server. If User
A comes to the message board, the address that will be logged
into the log file may look like this: www-ao2.proxy.aol.com.
The next time they log into the message board, their Internet
address may look like this: www-tj22.proxy.aol.com. It is impossible
to tell whether this user is a return visitor or another person
entirely. For this reason, I excluded America Online from my
calculations of return visitors, and made the assumption that
less popular ISPs with proxy servers represent return visitors.
America Online was the only major Internet service providers
(ISPs) that appears in the log files and uses proxy servers.
One of the eight participants of the chat room returned multiple
times. This person returned a total of seven times after the
first time. I talked with this person on the chat room the first
time he logged in, and we talked for about half an hour. One
of the disadvantages of the chat room is that the chat room is
synchronous, so people must be on at the same time to make the
chat room interesting. In addition, a certain critical mass of
participants is necessary. LambdaMOO, the most successful MOO
ever, has over 6000 users, and at any time there are usually at
least ten to twenty people on at a time. In retrospect, it seems
optimistic to expect that a group of much less than that could
reach a critical mass so that people would have an opportunity
to talk with each other often. It would take about 4000 people,
perhaps, to have two four-people conversations going on at once,
based LambdaMOO's rates.
Chat Room First Log-in Times
1.1 Synchronous and Asynchronous Group Communication
1.2 Research Questions
1.3 Outline of the Thesis
2. LITERATURE REVIEW
2.1 Groups: Identity, Formation, and Communication
2.2 Online Communities
2.3 Online vs. Face-to-Face Group Communication
2.4 Computer-Mediated Group Communication
2.41 Asynchronous Communication: Usenet
2.42 Asynchronous Communication: Online Mailing Lists
2.43 Synchronous Communication: MUDs and MOOs
2.44 Synchronous Communication: IRC and Chat Rooms
Figure 2-2Figure 2-2
3. PROGRAM DESCRIPTION
3.1 Chat Room - "SafeHouse"
3.2 Message Board
4. THE STUDY
4.1 Chat Room Program Implementation
4.2 Message Board Program Implementation
4.3 Data Collection: Research Questions
4.4 Data Collection: Questionnaire
5. RESULTS
5.1 Finding: Subjects Preferred to Use the Message Board
Figure 5-1Figure 5-1
5.2 Finding: Subjects Returned More to Message Board
