Content uploaded by Nicolas Ducheneaut
Author content
All content in this area was uploaded by Nicolas Ducheneaut
Content may be subject to copyright.
More than just ‘XP’: learning social skills in
massively multiplayer online games
1. INTRODUCTION
Playing computer games is becoming more and
more a social experience. Players often sit together
in front of a single machine, sharing skills and
expertise to accomplish a game’s objectives (King &
Borland, 2003); and as soon as machines could be
connected to one another, gamers were quick to
exploit the possibility to live and play in shared vir-
tual worlds (Cherny, 1999; Cuciz, 2001). It took the
recent explosion in Massively Multiplayer Online
Role-Playing Games (MMORPGs) – games like
Ultima Online,1EverQuest2and Star Wars
Galaxies3– however, for mainstream media and
public opinion to realize that computer games have
become full-fledged social worlds in their own right
(e.g. Kolbert, 2001).
Despite this new attention, little is know about
how these games work in practice as social worlds.
What kinds of interactions do players have with
each other? What kinds of knowledge and practices
are shared across the game community? To shed
more light on these issues we are conducting a
broad, exploratory study of the social dimensions of
multiplayer online games. As part of this study we
examine the range of social interactions that are
characteristic of MMORPGs and the opportunities
they provide for exercising and learning basic social
skills.
There has been a lot of research in the past on the
use and effects of games on teaching. Games that
are used in professional contexts include:
•Games for teaching a specific curriculum (e.g. mathe-
matics). Henry Jenkins and the members of the
Education Arcade symposium (see Terdiman,
2004), for instance, are interested in developing
or using games to teach students some of the
Interactive Technology & Smart Education (2005) 2: 89–100
© 2005 Troubador Publishing Ltd.
Nicolas Ducheneaut and Robert J. Moore
Palo Alto Research Center (PARC), 3333 Coyote Hill Road, Palo Alto, CA 94304, USA
Email: {nicolas,bobmoore}@parc.com
Massively Multiplayer Online Role Playing Games (MMORPGs) have become complex social worlds. As such, playing
these games requires more than accomplishing simple objectives: it is also a process of socialization into a communi-
ty of gamers. Through our observation of players’ activities we describe how MMORPGs provide opportunities for learn-
ing social skills such as: how to meet people; how to manage a small group; how to coordinate and cooperate with peo-
ple; and how to participate in sociable interaction with them. We show how this social learning is tied to three impor-
tant types of social interaction that are characteristic of MMORPGs: players’ self-organization, instrumental coordina-
tion, and downtime sociability. We conclude by discussing the societal impacts of our findings and how the features of
MMORPGs could be repurposed in environments specifically designed for social learning.
Keywords: Online games, ethnography, socialization, social behavior, social learning.
VOL 2 NO 2 MAY 2005 101
basic knowledge traditionally acquired in school.
While not directly concerned with games, the
work of Abelson and diSessa (1981) illustrates
how computers can be used in a “fun” way to
learn a curriculum.
•Games for therapy. Coming essentially from the
field of psychology, a number of past and current
efforts are under way to build games that have a
therapeutic component. The Self-Esteem
Games project at McGill University (Baldwin et
al., 2004), for instance, designed games to help
players feel more secure and confident in them-
selves.
•Games for training. These particular games are
used to put players in a specific situation repro-
ducing real-life. Flight simulators, such as the
popular version offered by Microsoft, are the
most well known examples.
Closer to our own research interests, games have
also been designed specifically to teach interpersonal
and intercultural communication skills (Raybourn,
2001). But these, as well as the other examples we
have mentioned before, all have a point in common:
they were designed consciously and specifically to promote
learning. In this they differ from most of the com-
puter games played nowadays, MMORPGs includ-
ed, which are designed primarily with entertain-
ment in mind. But while enjoyment is often the
principal benefit of playing video games, the more
recent crop of MMORPGs make social learning
part and parcel of the game (Jakobson & Taylor,
2003). Some game designers have clearly expressed
the intent to create games where socialization is
encouraged and rewarded (see Koster, 2004). As
these games have been quite successful at attracting
large numbers of players (Woodcock, 2004), they
could be an interesting source of design guidelines
for games that purposefully seek to encourage social
learning (Gee, 2003).
In this paper, we report on the initial findings
emerging out of our empirical studies of player-to-
player interactions in and around multiplayer
games. We document the cultural practices of
MMORPGs’ players, and how the process of social-
ization into a community of gamers is an important
source of learning. We focus on three major areas:
player self-organization, instrumental interactions,
and sociability. This allows us to show how
MMORPGs are about much more than just gaining
“experience points” (xp4). Instead, MMORPGs are
virtual platforms players can use to learn (and later
teach) the interpersonal communication principles
that are appropriate in online 3D environments. We
conclude by highlighting the particular features of
these games that directly contribute to social learn-
ing, and how they could be reused in other contexts.
2. BACKGROUND AND METHODS
In this study, we focused on one particular
MMORPG: Everquest Online Adventures (EQOA),5
an adaptation of the popular PC game Everquest to
Sony’s Playstation 2 (PS2) console. We selected this
game essentially for practical reasons: the two
authors both own a PS2, and capturing video data
from the game was facilitated by the fact that the
game was designed to be played on a standard tele-
vision, making its text easily readable when record-
ed on videotape. Moreover EQOA is a good repre-
sentative of the popular “medieval fantasy” genre of
MMORPGs, extending game design principles test-
ed in its predecessor Everquest but also earlier
games such as Ultima Online.
As of today, there are more than 50,000 people
playing EQOA in the United States (Woodcock,
2004). The game takes place in a fantasy world
broadly inspired from the works of authors such as
J.R.R. Tolkien. The game is set five centuries
before the PC-version of EverQuest and takes place
on a single continent, Tunaria, in the larger world of
Norrath. Much like pencil-and-paper role-playing
games, players select a “race” (e.g. elf, human, dwarf,
etc.) and a “class” (e.g. wizard, warrior), both of
which will affect their attributes and abilities.
Players then take control of an “avatar” or virtual
body in an elaborate 3D space, where they battle a
variety of creatures and accomplish quests to
progress in the game and develop their character
(see Figure 1).
To study life in EQOA we chose to adopt an
ethnographic stance. Ethnographic field research
involves the study of groups and people as they go
Ducheneaut and Moore: Learning Social Skills in Multiplayer Online Games
INTERACTIVE TECHNOLOGY & SMART EDUCATION
90
Figure 1 EQOA’s interface (above, players are forming a group)
Ducheneaut and Moore: Learning Social Skills in Multiplayer Online Games
about their everyday lives. The term “participant
observation” is often used to characterize this
approach, since the researcher seeks to immerse
himself in others’ worlds in order to grasp what they
experience as meaningful and important.
Ethnography therefore entails “some amount of
genuinely social interaction in the field with the
subjects of the study, some direct observation of rel-
evant events […] and open-endedness in the direc-
tion the study takes” (McCall & Simmons, 1969).
However, virtual worlds such as the ones repre-
sented by EQOA pose a methodological challenge
to the ethnographer. Indeed most of this approach
is based on the ethnographer “being there” in the
field to observe – but this “there” is nebulous at best
in the case of online spaces (Rutter & Smith, 2002).
Still, a majority of researchers believe in the virtues
of “virtual ethnography” (Hine, 2000), that is, an
adaptation of traditional ethnography to the study
of cyberspace. As Mason puts it:
“A virtual ethnography is one that fully immerses
the ethnographer into the consensual reality experi-
enced by groups of people who use computer-mediat-
ed communication as their primary, and often only,
means of communication. As such, the online or vir-
tual persona of the participants are the main focus
of the ethnographer. Generally, researchers have
wanted to focus on the person at the keyboard; a
virtual ethnography reverses this and works
instead with the persona that has been projected
into cyberspace by the typist.”
A virtual ethnography is then, simply, an ethnogra-
phy that treats cyberspace as the ethnographic real-
ity. This remains a controversial step, but it is not
the purpose of this paper to discuss this controver-
sy in great depth. The interested reader is referred
to (Hine, 2000; Lyman & Wakeford, 1999; Mason,
1999; Miller & Slater, 2000; Rutter & Smith, 2002)
to get an overview of the debate.
Our study started with the selection of a server
from one of the six EQOA offers in the United
States and the creation of new characters. To bal-
ance our view of the game as much as possible, one
of the authors selected a combat-oriented class
while the other selected a “combat-assist” class
(namely, a wizard – see section 3.4). We logged in
regularly – at least twice a week, sometimes much
more, each time for at least two hours over a three-
month period. We tended to play at the peak times,
which seemed to be weeknights, especially Fridays,
and afternoons on the weekend. Through this regu-
lar participation we progressively became members
of the community of players on our server. As our
characters evolved we joined “guilds” (semi-perma-
nent groups of players) and participated in activities
of increasing difficulty and complexity. All of these
activities were recorded using a video camera con-
nected directly to our consoles. This provided us
with a rich set of ethnographic data we later ana-
lyzed in depth using a combination Conversation
Analysis (Sacks et al., 1974) and open-ended coding
(Glaser, 1998). Overall we recorded close to 100
hours of gameplay from EQOA.
Through our participant observation of EQOA,
we developed a member’s perspective of what life is
like there. This perspective then greatly enhanced
our ability to understand the natural player prac-
tices captured on our videotapes. By reviewing the
videotapes and reflecting on our own experiences,
we began to identify the primary types of social
interaction in which players tend to engage – from
the initial “grouping sequences”, when several play-
ers decide to band together, to the humor and small
talk occurring between game events. Through these
observations we can see how players acquire the
skills needed to participate fully and meaningfully in
the wider community of a game.
3. FINDINGS
3.1 From “XP” to Social Capital
Early on in our observations, we were struck by the
social complexity of current multiplayer online
games such as EQOA. Most accounts of games
focus on the seemingly mindless task of killing mon-
sters and accumulating experience points to make
your character evolve. While it is true that players
spend a significant amount of time in combat
(which is not as mindless as it may seem – see
Taylor, 2003), they spend even more time simply
communicating with other players. At a minimum,
talking with others helps to accomplish the game’s
objectives (e.g. asking questions about the location
of an object or creature). But more often than not,
talking with others is an intricate part of the game.
For instance, Everquest’s “quests” are too difficult
for a single player so that only a coordinated and
complementary group of players can accomplish
them (Jakobson & Taylor, 2003): composing a group
and performing a specific role within it are there-
fore essential tasks in the game.
As a result, gamers need to do much more than
mindlessly accumulate experience points (xp): they
also need to increase their social capital within the
game’s society. In other words, they need not only
learn the game commands, but they must also
VOL 2 NO 2 MAY 2005 91
become socialized into the game community. To be
recognized as a good player you need to learn the
lingo, perform your instrumental role well when
grouped with others, and more generally demon-
strate that you are an interesting person to play with
(e.g. through humor). If you succeed, others will
include you in their “buddy list” to encourage fur-
ther interactions.7In short, these games are all
about having the right social skills.
3.2 Situated Learning in EQOA
As players enter the game world, they are confront-
ed by a bewildering array of new and foreign con-
cepts – much like a stranger entering a foreign cul-
ture. Online games, particularly of the role-playing
fantasy genre, share a rich culture that dates back to
at least the 1970s. This includes entities, concepts,
lingo and practices evolved out of the early text-
based Multi-User Dungeons (Cherny, 1999), pencil-
and-paper based games such as Dungeons and
Dragons (Fine, 1983; King & Borland, 2003), and
interactional practices borrowed from Internet
Relay Chat (IRC) and Instant Messaging (IM), for
example, brb or “be right back,” afk or “away from
keyboard,” and lol or “laugh out loud”.
The game manual, however, says nothing about
these concepts: instead, the greatest resource in
learning how to play is fellow players. In fact, it
seems game companies have even acknowledged
this fact implicitly. Game manuals are frequently
quite skinny, limiting themselves to a cursory
description of the most basic commands. Players
are encouraged to ask questions in the game and to
rely on the players’ community for knowledge (see
also Raybourn, 1998).
Although it is not an explicit goal of MMORPGs,
social learning nonetheless occurs all the time as a
normal feature of participation in a “community of
practice” (Lave & Wenger, 1991; Gee, 2003). In the
course of playing the game, there are a multitude of
opportunities for social interactions with other
players and through these “situated learning”
occurs. Opportunities for situated learning in the
MMORPG we have examined fall into four cate-
gories or modes of social learning:
•In-game, in-context discussions. The social nature of
MMORPGs is due in great part to their com-
munication infrastructure. All MMORPGs
offer, at a minimum, a textual chat system simi-
lar to IRC. It is used by the players to commu-
nicate whenever something of interest happens
in the game. This allows for a tight feedback
loop between the events happening on the
screen and the comments of the members of a
group. In other words, MMORPGs allow infor-
mation to be exchanged “on demand and just in
time, not out of the contexts of actual use” (Gee,
2003) – a central principle of good learning envi-
ronments.
•Out-game, out-of-context discussions. Communication
between the players does not stop when the game
is over. Instead, the game spills out into forums
and websites, some maintained by the game pro-
ducers, and some entirely player-created. These
discussions are also important occasions for
knowledge transfer. Players can be producers of
knowledge and not simply consumers (Gee, 2003).
•Observation. Most of the game’s activities are
accomplished in plain view of other players.
Players learn a great deal simply by watching
what other players do (Prensky, 2001). Observing
activities in a densely populated spot allows the
players to engage in “legitimate peripheral par-
ticipation” (Lave & Wenger, 1991).
•In-situ teaching. For many in-game practices, ask-
ing about them, reading about them, and observ-
ing them are not enough. Learning to do them
oneself requires practice and teaching from a
more experienced player. We found episodes of
direct teaching between players to be quite com-
mon (as in Raybourn, 1998).
In learning the game community’s shared culture
and practices, players draw on all four of these
modes of social learning.
Although Lave and Wenger (1991) point out the
centrality of practice in social learning, they say sur-
prising little about particular practices within which
learning can be situated. In the remainder of this
paper, we drill down on such shared practices in
EQOA. Furthermore, we highlight the potential
these practices provide for learning basic social
skills in addition to learning the particular game
practices themselves. In this vein we identify three
types of social interaction that are most important
for becoming a competent member of the game
community:
•Self-organization among players: as we mentioned
earlier, accomplishing the game’s objectives
requires coordination among players. For this
players must organize themselves into small
groups. Activities such as group creation, group
maintenance, and group disbanding are impor-
tant moments where the players can observe and
learn how to behave as a member of the game
community.
Ducheneaut and Moore: Learning Social Skills in Multiplayer Online Games
INTERACTIVE TECHNOLOGY & SMART EDUCATION
92
Ducheneaut and Moore: Learning Social Skills in Multiplayer Online Games
•Instrumental coordination: this covers moments in
the game when players have to work together as
a team to accomplish the game’s objectives.
Being able to perform a role reliably is a mark of
a player’s social competence.
•Sociability: this includes things such as humor (or
its absence), small talk, players “catching up”
with each other. This is as important to social
cohesion, if not more, than instrumental coordi-
nation.
We will now describe how social skills are acquired
and demonstrated (or not) during these three types
of interactions and four types of contexts. We will
illustrate how successful and unsuccessful social
learning happens through several example of player-
to-player interaction extracted from our video data.
3.3 Small-Group Self-Organization:
How to Meet People
Games like EQOA strongly encourage players to
team up with each other. The quests are specifical-
ly designed to be too difficult for a single character,
and so players must find help in order to complete
them. In addition, experience points can be earned
more efficiently if one is in a group with others.
Therefore, there is a built-in motivation to band
together with other players. However, there is no
automatic way to join a group (unlike in another
successful genre of multiplayer games: First-Person
Shooters. In games such as CounterStrike,8any new
player is automatically assigned to a team). In
EQOA players are on their own to find others who
are willing to play with them, and it is only through
interaction with other players that one can find will-
ing teammates.
Technically “grouping” simply involves one player
using the “invite to group” command and the recip-
ient selecting “yes.” Once grouped, the two share a
“groupchat” channel and automatically share experi-
ence points from enemies killed by either player.
The most efficient way to gain experience points is
to battle opponents while grouped with others. By
working as a team, you can defeat higher-level ene-
mies (which give more experience points) than you
could defeat on your own. The initial inviter is then
the “leader” of the group, in a technical sense, in
that only he or she can invite additional members
(four is the maximum group size) or disband the
group.
While technically grouping is quite simple, social-
ly it is more complex. Although any player can
extend a formal invitation to any other at any time,
there are standard shared ways of grouping. These
grouping practices emerge and become established
over time, and novice players, or “newbies,” learn
them through observation and participation. For
example, players tend to form groups in particular
places. In EQOA there is a “coach” system of trans-
portation by which characters can teleport between
towns. To use a coach, the character must go to a
stable in a town and talk to the coach master. It is
these coaches that informally serve as grouping
spots (see Figure 2). Players go to coaches to meet
up with other players looking to join a group, check
each other out, and offer invitations.
Once at a coach, there are a variety of ways to
form a group, yet the standard method is for one
player to solicit an invitation and another player
then to extend one. Players can use a “looking for
group” command which when used automatically
produces the following kind of message in the chat
box:
G shouts: 22 Shadowknight seeking group!
The format of the message accomplishes several
things: it displays the message to everyone within
“shouting” range, announces the level and class of
the character, as well as the desire to join a group.
Another player can assess the level and class of the
character to determine if he or she would be a good
fit and then approach the character, or more com-
monly, simply extend an invitation based on this
information.
Although the “looking for group” command is
part of the game’s design, players then improvise on
its basic format. That is, players will manually type
(or program) customized messages that resemble
the “looking for group” message, but which provide
more fine-grained information. For example:
VOL 2 NO 2 MAY 2005 93
Figure 2 Grouping at coach
B shouts: Lvl 29 SK Seeking XP Group!
O shouts: Level 23 Frontier Cannibal Shaman
needs group for Long XP Grind
V shouts: A lvl 9 bard seeking a xp grp…PST
or INVITE
M shouts: UNEMPLOYED 13 Shaman
LOOKING 4 A GROUP W/ A REAL
TANK!
L shouts: Group around level 23 with a healer
looking for more!
M shouts: Grp seeks a caster lvl 20-23 PST9
In these variations on the “looking for group” mes-
sage, players may reveal more about their charac-
ters, reveal what kind of group they wish to join,
and reveal what they want the group to do.
Other important aspect of grouping practices are
level matching and class balancing. In order to fight effec-
tively and maximize experience points, a good group
should contain four members, all around the same level
of experience (levels range from 1 to 60), and with com-
plementary classes (combat roles are explained in the
next section). A good group leader will assemble the
most balanced group possible given the characters
available on the scene. When deciding whether to
invite a character or whether to stick with a group,
players will assess whether the group is adequately bal-
anced. For example, in the following exchange, E, C
and M are in a group and have just lost their fourth
member. They return to the coach at Darvar Manor to
recruit another player. E, the group leader, uses the
“Who” menu to scan the names and attributes of all
characters in the vicinity. The members then discuss
what class of character to try to enlist.
C tells the group: what should we get… cast-
er? tank?
C tells the group: these ogre shadowknights
look badass
M tells the group: get damage deal if u can
E tells the group: what’s damage deal?
M tells the group: a damage dealing class
In other words, M advises E to try to enlist a war-
rior-class character or “damage dealing class” (e.g.
warrior, paladin, shadowknight) who can inflict
extensive damage to an opponent using weapons.
In another case, R is assembling a group for the
explicit purpose of traveling to a particular town,
Moradhim, rather than for the typical purpose of
gaining experience points. He also specifies the
range of levels he is looking for, which is rather large.
R shouts: looking for a group lvl 8 to 12
R shouts: for help getting to mor10
R (level 8) has successfully recruited A (level 9) and
then invites T (level 12), who is a significantly high-
er level than the others. When T accepts, A points
out that the inclusion of T in the group will pre-
clude A and R from gaining xp (based on the game
design) because T’s level is too high.
A tells the group: u know we wont get any xp now
A tells the group: no offense
R tells the group: i know and i don’t care for
it right now
It is through this interaction with A and R that T
learned that level matching matters when assem-
bling a group.
Usually after these formal invitations have been
accepted (although also sometimes before), the
players exchange bows and textual greetings and
begin to plan what they will do and where they will
go. At any time following the initial grouping, they
may discover that they do not all work well togeth-
er, that they have diverging plans, or even that they
do not like each other. A group member may then
leave, or be kicked out and begin the grouping
process all over again.
Thus, we see that grouping practices in EQOA
are somewhat complicated and that becoming com-
petent in them requires in-game socialization.
While the mechanics of how grouping is accom-
plished in the game is quite different than meeting
people in real life, in many ways they are similar.
The closest real-world equivalents to grouping at
coaches in EQOA may be situations like pick-up
basketball or singles’ bars. One goes to a particular
place, surveys the potential teammates or partners
who are available on the scene, approaches those
who seem to be the best fit, and in the end makes
due with whomever they can get.
In summary, grouping in EQOA gives players
experience in approaching and meeting strangers. In
some ways, virtual worlds are ideal places for learn-
ing to meet new people because they are safer than
real life and the costs of rejection and losing face
seem to be much lower. In this they resemble sin-
gles’ bars, or pick-up sports games. They also teach
how to assemble a well-balanced, efficient team – an
important skill in today’s workplace (Gee, 2003).
3.4 Instrumental Coordination: How to
Coordinate and Cooperate with
People
Once a group is formed, each player has to learn
how to play his or her role correctly: it is not just
Ducheneaut and Moore: Learning Social Skills in Multiplayer Online Games
INTERACTIVE TECHNOLOGY & SMART EDUCATION
94
Ducheneaut and Moore: Learning Social Skills in Multiplayer Online Games
hacking and slashing. As we mentioned earlier the
game’s quests are purposefully too difficult for a sin-
gle player, and group members must coordinate
their actions in order to succeed. Over time, the
entire community of players has evolved several cul-
tural practices to deal with this issue of coordina-
tion, practices that newcomers must learn.
Initially when newbies engage in group combat,
they tend to attack the computer-controlled oppo-
nent in individualistic ways. Each group member
does their own thing, but they are not necessarily
coordinating their attacks. However, when playing
with more experienced players (or by browsing the
game message boards), they soon begin to hear men-
tion of particular combat roles: “tank”, “healer,”
“caster,” and “melee.” A good group is one that is
balanced in terms of the classes that can play these
roles, and experienced players know how to play the
role that is appropriate to their character’s class.
The most efficient combat strategy is to attack
only one opponent at a time. This is the responsi-
bility of the tank (rugged weapon-fighting classes):
he “pulls” (or attracts) a single opponent toward the
group, and then “taunts” it so that it only attacks
him (see G in Figure 3). This facilitates the work of
the other group members. While the opponent is
focused on the tank, the healer (see C in Figure 3)
can then focus his efforts on just one person – the
tank – although the healer is responsible for keep-
ing everyone in the group from dying. Failure can
occur if the tank pulls several opponents at once,
causing the healer to be overwhelmed, run out of
power, and eventually everybody dying for lack of
healing. Similarly, the tank also distracts the oppo-
nent from the caster. The caster (spell-casting class-
es; see character in foreground in Figure 3) can
cause the greatest amount of damage to an oppo-
nent using spells, yet he is extremely vulnerable and
will die quickly if attacked. It is the responsibility of
the tank to make sure this doesn’t happen and, if it
does, the responsibility of the healer to remedy it.
Finally the melee (weaker weapon-fighting classes;
see M in Figure 3, who is barely visible between the
gnoll and the tree) assists the tank in combat and in
keeping opponents away from the healer and caster.
In other words, these roles, and the corresponding
classes, are interdependent: individually each one is
fairly limited, but when working together, they can
be more effective than any of them could be on
their own. In playing these four combat roles, if a
group of players can tightly coordinate their
actions, they can efficiently eliminate hordes of
opponents. The result is an “xp grind” which
enables them to progress through the levels of their
class at a rapid pace.
While learning the terminology and functions of
the combat roles is a significant accomplishment
for the newbie, these are not enough. Performing a
particular role takes skill and practice. Within a
group, a great deal of learning goes on. More expe-
rienced players often teach “newbies” the sub-
tleties of their role by offering tips during the
course of combat or modeling correct technique.
For example, on one occasion, C and E are playing
with a much more experienced Z who is “power-
leveling” them. In power-leveling, a more advanced
character helps less advanced characters gain xp at
an accelerated rate. Although the designers have
tried to prevent any type of power-leveling, the
players have discovered a rather complicated
workaround. This workaround has become part of
the shared game culture, and on this occasion, Z is
teaching the others how to do it. C is learning to
play the role of “healer” and E “caster.” Z advises
them on each role:
Z: are you chaining heals or just using one?
C: I’m casting it several times, should I be
doing something different?
Z: use two different heals and alternate
E: z, what are you casting?
E: or were you casting during the fight?
Z: detonate once then on second it was freez-
ing strike
Z: i didnt till mob was half down or youll lose xp
E: what are u going to cast to keep the
E: nm
Z: I cant keep agro
E: right
Z: or ill do too much damage. I can only drop
once you get it half way
E: so you’re just doing less than half the dmg?
E: got it
VOL 2 NO 2 MAY 2005 95
Figure 3 Combat roles
For the novice player, power-leveling is rather
confusing; however, through player-to-player teach-
ing, it can be mastered after only a few sessions. As
illustrated above it requires the low level player to
start attacking a single monster (or “mob”) until its
health is down to half its initial value, at which point
the higher level player steps in to kill the monster in
one blow (here, using a powerful spell such as “det-
onate” or “freezing strike”). Any breakdown in this
process can result in players not receiving experi-
ence points, dying, or both. This example illustrates
how, more generally, learning to play a combat role
is a central element in fitting into the game com-
munity, and how it can only be learned in-game
through participation and interaction with other
players.
Playing EQOA is, therefore, a really important
source of learning how to be a good teammate.
Playing a combat role in the context of a group is
very similar to playing a particular position in a
sport in the real world. To be successful, players
must learn the technical aspects of coordinating
their actions, as well as the more general ability to
cooperate. In fact, our observations reveal that play-
ers who refuse (consciously or not) to play these
group roles are quickly shunned. Among the social
skills learned we can list:
• Leadership: advanced players must learn to be
good conductors and orient the activities of new-
comers. Style is of the essence here: this is a
game after all, in which players want to have fun.
As such, authoritarian dictators are rarely suc-
cessful. Instead, good high-level players are more
empathic. They reinforce “good” group behavior
(e.g. a tank carefully pulling a single monster) and
point at coordination problems (e.g. recommend
that a caster stands away from combat).
• Sensitivity to others’ needs: players need to
observe the activities of other group members
and adapt their actions accordingly. If the healer
is overwhelmed and out of power, for instance, it
is bad practice for the tank to immediately start
bringing fresh monsters toward the group.
Players who do not act according to others’
needs are quickly excluded from a group.
3.5 Sociability: How to Socialize with
People
Combat sequences are only a fraction of the game.
EQOA has a certain rhythm built into it, where
combat alternates with periods of “downtime” (e.g.
between opponents or while traveling from one
location to the next). These calmer periods are an
opportunity for the players to chat with each other
about a variety of topics, game related or not. Some
game designers, such as Raph Koster of Ultima
Online and Star Wars Galaxies, consciously and
strategically use periods of downtime – short peri-
ods in which player must wait – in order to try to
encourage social interaction among players (Koster,
2004). Thus downtime is an important opportunity
for players to learn and exercise sociability (Simmel,
1949), that is, social interaction that is pursued for
its own enjoyment and need produce no extrinsic
results.
Indeed, playing MMORPGs is essentially about
hanging out with people recreationally. In some
ways, they simulate features of “third places”
(Oldenburg, 1989) such as a local pub: instead of
having a few drinks, a game of darts and a lot of
laughs with your friends, you battle a few monsters,
explore a rich landscape and have a lot of laughs
with your friends (for a deeper discussion of this
particular aspect of MMORPGs see Ducheneaut
and Moore, 2004). Games like EQOA are not
focused purely on instrumental coordination (i.e.
how to kill monsters): there are opportunities for
sociable interaction too.
As a lot of game time is spent “doing nothing” (or,
to put it differently, not killing anything) it is
important for a player to demonstrate that he is an
interesting person to be with. While some may
enjoy hours of running silently in the game’s wilder-
ness, we found that most players would much rather
use this time to talk to each other. During these
moments, humor and laughter (usually expressed
with “lol,” which is short for “laugh out loud”) are
extremely important for group cohesion and suc-
cess. For example, the following exchange occurred
while a group of friends was traveling to a dungeon
for a quest. All three players are male, but E is a
female character:
A tells the group: my wife is jealous that I am
playing with another woman wearing a sexy
robe
E tells the group: did u tell her about me?
E tells the group: us?
C tells the group: lol
A tells the group: yes, she saw the picture of u
I have in my wallet
E tells the group: LOL
On another occasion, a group has just formed at a
popular coach and the new group is now running to
a particular location for an “xp grind” or long, effi-
cient combat session. C and E already know each
Ducheneaut and Moore: Learning Social Skills in Multiplayer Online Games
INTERACTIVE TECHNOLOGY & SMART EDUCATION
96
Ducheneaut and Moore: Learning Social Skills in Multiplayer Online Games
other, but otherwise the group consists of strangers:
G tells the group: How long evryone on 4
C tells the group: not sure...
G tells the group: k
E tells the group: at least an hour
G tells the group: boo
M tells the group: im not sure yet
C tells the group: lol
E tells the group: depends on how fun you
guys are ;-)
G tells the group: lol
M tells the group: nice
In addition, humor can also be woven into the activ-
ity of grinding combat xp. For example, during a
long power-leveling session, Z and E who are both
wizards and thus can stand back a bit from the
action have the following exchange in which E asks
Z about an object his avatar is visibly holding:
E: oo, i like your book, z
Z: its a list of ex wives
E: lol
While good jokes can make the gaming experience
more enjoyable for everyone in a group, failed jokes
can detract from it. A player who can make the
other players in a group laugh is highly valued and
may be kept around despite inexperience in com-
bat. However a player who repeatedly tells jokes
that the rest of the group does not find funny will
likely be kicked from the group. For example, on
one occasion, three friends, E, A and C picked up a
fourth player, B. It quickly became clear that B did
not fit in with the rest of the group who liked to tell
numerous jokes. However, when B tries to partici-
pate in the joking, he is unsuccessful (see Figure 4).
E tells the group: the plan is to power level
C tells the group: cool
B tells the group: i have a 17 mag that could
pop your head like a zit
((long pause))
B tells the group: lol
((longer pause))
E tells the group: the question is where and
what color should the mob be
Eventually the group abandoned B, in part, because
he simply was not funny.
Periods of downtime are thus a perfect platform
for players to experiment and learn about sociable
behavior – a skill that could easily translate to the
physical world. Players learn about how and when to
use humor, and how to approach strangers and pro-
gressively build up relationships. Moreover, games
are an ideal platform for experimenting with socia-
bility for two reasons:
• There is always something to talk about. The
game’s objectives are ideal conversation starters
and ice-breakers: players share a lot of common
ground because of the game’s framework – if
nothing else, you can always talk about the last
monster killed or where to go next. This pre-
vents what we call “interactional paralysis”:
unlike other physical or virtual spaces, the con-
text of the game encourages interaction.
MMORPGs benefit from a kind of “triangula-
tion” similar to the one discussed by Whyte
(1988, p.154).
• The mediation of a virtual avatar, the use of
pseudonyms, and text-based communication all
reduce the risks of failed interactions. Unlike
real life, there is little stigma for experimenting
with new jokes or trying to approach unknown
others. As such games are interesting platforms
for testing interactional strategies that can later
be reused (or not) in the physical world.
4. CONCLUSION
There is little doubt that MMORPGs have become
complex social spaces. Playing them is about more
than mindlessly killing monsters: it is about learning
and participating in the shared practices of a game
community. As we argue, such online games provide
opportunities for learning social skills such as how
to meet people, how to manage a small group, how
to coordinate and cooperate with people, and how
VOL 2 NO 2 MAY 2005 97
Figure 4 A failed joke
to participate in sociable interaction with them.
Through our virtual ethnography, we have begun to
describe the types of social interactions—self-
organization into small groups, instrumental coordi-
nation, and sociability – that are characteristic of
MMORPGs such as EQOA and that result from
features of their design (see Table 1).
Our description of the types of social interaction
in MMORPGs and the basic social skills required
to participate in them raises the question of how
these skills may carry over into real life or not.
Online social worlds such as MMORPGs provide a
relatively “safe” environment to experiment with
interpersonal communication. Game characters
provide players with a great deal of anonymity; the
pace of interaction is slowed by the use of text chat
giving players more time to think about responses;
offending another player cannot escalate into real
violence and textually abusive players can be
silenced through the “ignore” function. Hence these
online worlds are relatively safe places to experi-
ment with approaching others and proposing joint
activity, with playing specific roles within a small
group, with initiating personal talk, joking, and
more. Through practice in these virtual environ-
ments, players may build up self confidence in inter-
acting with and organizing others that could carry
over into real life. In fact, one veteran player of
Everquest and Star Wars Galaxies whom we inter-
viewed told us that although he is “pretty shy” in
real life, online he is a leader in his guild (an in-game
club or player association) and he enjoys spending
time helping “newbies.” He asserted that he
thought the several years he spent playing
MMORPGs have made him more outgoing socially
in real life. In fact there is already some evidence
that playing MMORPGs can improve one’s basic
social skills. When asked about whether they
thought game play had improved their leadership
skills in real life, roughly half of the Everquest play-
ers surveyed reported that they had learned a little
or a lot in terms of four leadership skills: mediation,
persuasion, motivation, and overall leadership (Yee,
2004). Other research suggests that Collaborative
Virtual Environments (CVE) similar to
MMORPGs are useful for teaching basic social
skills to adults with Asperger’s Syndrome (Parsons
et al, 2000; Beardon et al, 2001). “The shared fea-
tures between virtual and real worlds may facilitate
the generalization of skills from the former to the
latter” (Kerr 2002: 81). Certainly more empirical
studies must be done in order to measure the
impact of playing MMORPGs on real-life social
competences.
Moreover the design of MMORPGs fosters the
development of social skills by encouraging players
to interact with one another. This is done primarily
by doing three things: 1) creating tasks, such as
quests, that require the participation of multiple
players to accomplish (discouraging “solo play”), 2)
creating interdependencies between the different
character “classes” and combat roles and 3) building
periods of downtime into the rhythm of game play.
These seem to be general principles of small group
dynamics which should be effective in a wider vari-
ety of CVEs beyond MMORPGs. There is no rea-
son that the quests and character professions must
be combat-oriented and set in fantasy worlds. In
fact, even within the MMORPG genre itself there
is beginning to be a wider range of activities and
professions. For example, in the game Star Wars
Galaxies, character professions include doctors,
dancers, chefs, architects, mayors, and more that
are non-combat-oriented. Players who choose these
professions never have to engage in combat in order
to progress in the game. Thus we can imagine
Ducheneaut and Moore: Learning Social Skills in Multiplayer Online Games
INTERACTIVE TECHNOLOGY & SMART EDUCATION
98
Small-group self-organization Instrumental coordination Sociability
Activities • Soliciting invitations with “looking for • Pointing out a potential target • Initiating small talk
group” announcements to the group • Telling jokes
• Identifying best potential group members • Coordinating combat actions • Being sensitive to personal
• Inviting individuals to join the group with other group members talk
• Looting defeated opponents
Game knowledge • Level matching • Combat roles: tank, healer, • Chat commands
• Class balancing caster or melee
• Grouping commands • Pulling
• Powerleveling
• Combat commands
Basic social skills • Approaching strangers • Cooperation • Politeness
• Recruiting • Playing team role • Humor, wit
• Small-group formation • Team coordination • Empathy
Table 1 Activities, knowledge and basic social skills by types of social interaction
education-oriented worlds that use the same princi-
ples of social interaction but use different content
(perhaps based on history or real-world profes-
sions).
Going a step further the experience-points-based
achievement systems in MMORPGs could easily be
transformed into educational-credits-based
achievement systems in which students accumulate
credits for accomplishing educational tasks. Instead
of “grinding” through virtual combat, students
could grind through assignments together. In order
to motivate advancement, students would gain
some kind of new ability for their in-game charac-
ters whenever they reach a new level. Of course
another motivation is reputation among fellow stu-
dents. Thus the structure of MMORPGs may be
suitable for learning some level of educational cur-
ricula in addition to learning basic social skills.
Overall, our research highlights the positive
impacts MMORPGs can have on their players. As a
final caveat, it is worth mentioning that all interac-
tions in a multiplayer game are not necessarily the
source of social learning – for instance, a class of
players known as “griefers” engages in activities
with the specific intent of disrupting other players’
enjoyment of the game (Foo, 2004). In our experi-
ence however, these incidents are quite rare. In the
words of Baym (1995), it is easy to focus on these
problematic and highly publicized incidents and
loose sight of the “countless, rewarding and routine
non-problematic interactions” taking place in mul-
tiplayer games. Like Bartle (2004), we believe
instead that “virtual worlds are a force for good,
based on the fact that players can learn to be better
people as a result of playing them.” Much can be
gained by understanding how MMORPGs can
make social learning such an enjoyable activity.
NOTES
1. http://www.uo.com/ (January 2005).
2. http://eqlive.station.sony.com/ (January 2005).
3. http://starwarsgalaxies.station.sony.com/ (January
2005).
4. Gaining experience points is a central part of playing
MMORPGs. Earning them is the only way for players
to progress through the levels of their character’s “class”
(e.g., wizard or warrior) and, in return, to gain more
powerful abilities (e.g. stronger spells, deadlier attacks).
5. http://everquestonlineadventures.station.sony.com/
(January 2005).
6. To protect the players’ privacy, all names have been
deleted from our figures.
7. Eventually groups of players who enjoy their time
together can form “guilds” or “clans”, which are semi-
permanent social structures for organizing their
members’ activities. As a member of a guild, a player
can more easily find other people to play with and can
participate in “guild raids” in which dozens of guild
mates work together to defeat an especially difficult
opponent such as a dragon. To achieve this, guild
leaders often pay close attention to the balance of
classes and skills offered by its members.
8. http://www.counter-strike.net/ (January 2005).
9. Lvl: level; SK: ShadowKnight (a character class); Grp:
group; PST: Please Send Tell (that is, the player is
asking others to reply using the private messaging
channel).
10. Mor means Moradhim – a centrally located village in
the game world.
REFERENCES
Abelson, H. and diSessa, A. (1981) Turtle geometry: The
computer as a medium for exploring mathematics.
Cambridge, MA: MIT Press.
Baldwin, M., Baccus, J., Dandeneau, S. and
Sakellaropoulo, M. (2004) Self-esteem games. From
http://www.selfesteemgames.mcgill.ca/ (January 2005).
Bartle, R. (2004) Designing virtual worlds. Indianapolis,
IN: New Riders.
Baym, N.K. (1995) From practice to culture on Usenet.
In S. L. Star (ed.), The cultures of computing. Cambridge,
MA: Blackwell, pp. 28–52.
Beardon, L., Parsons, S. and Neale, H. (2001) An inter-
disciplinary approach to investigating the use of vir-
tual reality environments for people with Asperger
syndrome. Educational and Child Psychology, 18(2), pp.
53–62.
Cherny, L. (1999) Conversation and community: Chat in a
virtual world. Palo Alto, CA: CSLI Publications.
Cuciz, D. (2001) The history of MUDs. From
http://www.gamespy.com/articles/january01/muds1/in
dex4.shtm (January 2005).
Ducheneaut, N. and Moore, R.J. (2004) Designing for
sociability in massively multiplayer games: An examina-
tion of the ‘third places’ of SWG. In J.H. Smith and M.
Sicart (eds.), Proceedings of the Other Players conference
(http://www.itu.dk/op/proceedings.htm). Copenhagen:
IT University of Copenhagen.
Fine, G.A. (1983) Shared fantasy: Role-playing games as social
worlds. Chicago, IL: The University of Chicago Press.
Foo, C.Y. (2004) Redefining grief play. In J.H. Smith and
M. Sicart (eds.), Proceedings of the Other Players conference
(http://www.itu.dk/op/proceedings.htm). Copenhagen:
IT University of Copenhagen.
Gee, J.P. (2003) What video games have to teach us about
learning and literacy. New York: Palgrave/MacMillan.
Glaser, B.G. (1998) Doing grounded theory: issues and discus-
sions. Mill Valley, CA: Sociology Press.
Hine, C. (2000) Virtual ethnography. Thousand Oaks,
CA: Sage.
Jakobson, M. and Taylor, T.L. (2003) The Sopranos
meets EverQuest: Social networking in massively
multiplayer online games. Conference proceedings of the
2003 Digital Arts and Culture (DAC) conference.
Melbourne, Australia. pp. 81–90.
Kerr, S.J. (2002) Scaffolding – Design issues in single and
collaborative virtual environments for social skills
learning. In S. Muller and W. Sturzlinger (eds.),
Ducheneaut and Moore: Learning Social Skills in Multiplayer Online Games
VOL 2 NO 2 MAY 2005 99
Proceedings of the Eighth Eurographics Workshop on
Virtual Environments. Aire-la-Ville, Switzerland:
Eurographics Association.
King, B. and Borland, J. (2003) Dungeons and dreamers: The
rise of computer game culture. New York: McGraw-Hill.
Kolbert, E. (2001) Pimps and dragons. The New Yorker, May
28, 2001. From http://www.newyorker.com/fact/con-
tent/?010528fa_FACT (January 2005).
Koster, R. (2004) Raph Koster's writings on game design. From
http://www.legendmud.org/raph/gaming/index.html
(January 2005).
Lave, J. and Wenger, E. (1991) Situated learning: Legitimate
peripheral participation. New York, NY: Cambridge
University.
Lyman, P. and Wakeford, N. (eds.) (1999) Analyzing vir-
tual societies: New directions in methodology. Thousand
Oaks: Sage.
McCall, G. and Simmons, J. (1969) Issues in participant
observation. New York: Addison Wesley.
Mason, B. (1999) Issues in virtual ethnography. In K.
Buckner (ed.), Proceedings of the Esprit i3 workshop on
ethnographic studies. Edinburgh: Queen Margaret
College, pp. 61–69.
Miller, D. and Slater, D. (2000) The Internet: An ethno-
graphic approach. Oxford: Berg Publishers.
Oldenburg, R. (1989) The great good place. New York, NY:
Marlowe & Co.
Parsons, S., Beardon, L., Neale, H.R., Reynard, G.,
Eastgate, R., Wilson, J.R., Cobb, S.V.G., Benford,
S.D., Mitchell, P. and Hopkins, E. (2000)
Development of social skills amongst adults with
Asperger's Syndrome using virtual environments: The
'AS Interactive' project. In P. Sharkey, A. Cesarani, L.
Pugnetti and A Rizzo (eds.), 3rd ICDVRAT, Sardina,
Italy, pp. 163–170.
Prensky, M. (2001) Digital game-based learning. McGraw-
Hill.
Raybourn, E. (1998) The quest for power, popularity, and
privilege in cyberspace: Identity construction in a text-
based multi-user virtual reality. Paper presented at the
Western Speech Communication Association Conference,
Denver, Colorado, February 1998. Available at:
http://www.cs.unm.edu/~raybourn/moo5d~1.htm (January
2005).
Raybourn, E. (2001) Designing an emergent culture of
negotiation in collaborative virtual communities: The
DomeCityMOO simulation. In E. Churchill, D.
Snowden and A. Munro (eds.), Collaborative virtual
environments: Digital places and spaces for interaction.
London, UK: Springer Verlag, pp. 247–264.
Rutter, J. and Smith, G. (2002) Ethnographic presence in
a nebulous setting. In Hine, C. (ed.), Virtual methods:
Issues in social research on the Internet. Berg.
Sacks, H., Schegloff, E.A. and Jefferson, G. (1974) A sim-
plest systematics for the organization of turn-taking
for conversation. Language, 50, pp. 696–735.
Simmel, G. (1949) The sociology of sociability. American
Journal of Sociology, 55, pp. 254–261.
Taylor, T.L. (2003) Power gamers just want to have fun?:
Instrumental play in a MMOG. Proceedings of the 1st Digra
conference: Level Up. The University of Utrecht, The
Netherlands. Available at: http://www.itu.dk/people/tltay-
lor/papers/Taylor-Powergamers.pdf (January 2005).
Terdiman, D. (2004) Play games, be better students? From
http://www.wired.com/news/games/0,2101,63415,00.h
tml (January 2005).
Whyte, W.H. (1988) City: Rediscovering the center. New
York: Anchor Books.
Woodcock, B. (2004) An Analysis of MMOG
Subscription Growth – Version 9.0. From http://pw1.net-
com.com/~sirbruce/Subscriptions.html (January 2005).
Yee, N. (2004) Learning Leadership Skills. The Daedalus
Project. From http://www.nickyee.com/daedalus/
archives/000338.php (January 2005).
INTERACTIVE TECHNOLOGY & SMART EDUCATION
100
Ducheneaut and Moore: Learning Social Skills in Multiplayer Online Games
Nicolas Ducheneaut is a social scientist in the Computer Science Laboratory at PARC. He obtained his PhD in 2003 from the University
of California, Berkeley. His research interests include the sociology of online communities, computer-supported cooperative work, and
human-computer interaction. He has studied social interactions in a wide variety of environments ranging from real-time multiplayer
games to large bureaucracies. These studies led to the design of novel technologies to better support electronic communication in vir-
tual environments. Nicolas's research is based on a combination of qualitative methods (ethnography) with quantitative data collec-
tion and analysis (such as data mining and social network analysis).
Robert J. Moore is a sociologist in the Systems and Practices Laboratory at PARC. He obtained his PhD in 1999 from Indiana University,
Bloomington, where he was trained in ethnomethodology and conversation analysis. His research interests include the organization
of social interaction and practice in workplaces and in 3D virtual environments, with a focus on gesture and embodied conduct. He has
conducted video-based, work practice ethnographies of document production centers, multiplayer online games, automobile assem-
bly plants, and survey research call centers.
