Vol. 7, No. 1 (2013)





Played and Designed Sociality in a Massive Multiplayer Online Game

Lina Eklund, Magnus Johansson

Eludamos. Journal for Computer Game Culture. 2013; 7 (1), pp. 35-54



Played and Designed Sociality in a Massive Multiplayer Online Game

Lina Eklund and Magnus Johansson

1. Introduction

Online games are social spaces, new social worlds where players spend time and interact with others as well as form friendships or meet future life partners (Taylor 2006). The social play is the unique selling point of these games (Cole and Griffiths 2007) and design for interaction between players is a basic function (Williams 2006). Player interdependency, player created guilds, player grouping and other social engineering features are ways in which developers foster social engagement within the game and support player interaction (Jakobsson and Taylor 2003; Ducheneaut, Yee, Nickell and Moore 2006). In Massive Multiplayer Online Games (MMOs) players compete but also work together and have the opportunity of building long-lasting relationships; in short, these games offer many opportunities for social interaction (Kolo and Bauer 2004; Ducheneaut and Moore 2004). As Jakobsson and Taylor (2003, p.88) argue, "The production of social networks and the circulation of social capital prove to be one of the most important aspects in EQ" [The MMO EverQuest (Sony 1999)]. Gameplay where groups take part in different adventures is what characterises these games, so they are dependent on high levels of trust and cooperation between gamers in order to function (Chen 2008). Gamers do, however, use these spaces in their own ways and engage in sociality on their own terms (Simon, Boudreau and Silverman 2009). In this study we take a closer look at the relationship between played and designed sociality in online games from an interaction perspective by looking at sociality in temporary collaboration groups. A design change in the game World of Warcraft (WoW. Blizzard, 2004) is used as a case study. The introduction of the Dungeon Finder tool, has radically changed the premises for temporary collaboration groups and this alteration gives us a unique opportunity to study sociality and the relationship between designed and played sociality as this relationship changes.

Research on digital games has shown that MMOs can be viewed as 'third places'[1] (Ducheneaut, Moore and Nickell 2004), that is, places for informal sociability-sociability being one form of social interaction, which is defined by equality. We are all equals in a sociable situation of 'shallowness'; where personal discussions are kept out in favour of light-hearted interaction. MMOs are said to offer the opportunity for bridging social capital, connecting people who would not have known each other otherwise (Steinkuehler and Williams 2006). These social properties are an important reason for studying digital games (Williams 2006). Even instrumental players focusing more on personal advancement than the social aspects of these worlds are highly social, since group play is often the only way to progress in these games (Taylor 2003).

In MMOs, temporary collaboration or so-called pickup groups (PUGs), constitute a basic arena where players meet and join forces to take on greater challenges together than they could possibly have been able to do on their own (Myers 2005; Taylor 2006). In a PUG, players collaborate with people they do not know beforehand, and creating these-often short-lived-groups is a prominent element in many MMOs. Looking at group formation in MMOs, research has mainly focused on Guilds, that is, more permanent player groups (see e.g. Ducheneaut et al. 2006; Williams, Ducheneaut, Xiong, Zhang, Yee and Nickell 2006), while little attention has been directed towards temporary groups. We believe that studying these latter offer important insights into online collaboration, social worlds of MMOs and how design and practise are connected. Our specific research questions are: How can we understand social play in temporary groups? How do played sociality and designed sociality interact in these groups?

The chosen venue for this study is temporary collaboration groups or PUGs in World of Warcraft, released by Blizzard in 2004 and one of the dominating titles in the MMO genre. In order to explore sociality in this context, the study starts with interaction data gathered in temporary WoW groups, followed by a closer look at design, players' experiences and the gaming community.

Our arguments will proceed as follows. In the first part of the article, we outline the area of study. We then discuss the design of pickup groups in WoW and how this design has changed over time. This is followed by a brief theoretical discussion. In the second part of the article, we present the methodology and the data used, followed by results. The article ends with conclusions and a discussion of the relationship between design, player experiences and social gameplay.


2. Social gameplay

Simon et al. (2009) define two aspects of sociality in gaming: 1) Designed sociality, i.e. the social architecture/structure of the game; and 2) Played sociality, i.e. what gamers do within these structures. These two parts of sociality regulate social gameplay and are intertwined and dependent on each other. The way we use the terms here is that designed sociality is the affordances for social play within the architecture of the game while played sociality is the way gamers use and experience these affordances. Yee (2009), looking at computer mediated communication (CMC) in EverQuest (EQ), argues that designed sociality matters. EQ is a difficult game to play alone since the different game characters are highly dependent on each other. This dependency fosters a culture of seeking and providing assistance, while the social architecture provides ways for gamers to help others. The social architecture of EQ is partly therefore a manner of social engineering (Yee 2009) allowing certain game actions. However, in view of Simon et al. (2009), played sociality is an important part of online social life; gamers appropriate and use games to suit their own needs and wants.

The intention of this study is to enrich our knowledge of social interaction in online play, as well as to show how attention paid to play (designed and experienced) can give us valuable insights into these social worlds. If we want to study what happens when people interact with technology, here digital games, we need to take usage into consideration as well as how gamers make sense of their activities (Livingstone 2002). This can give us valuable insights into understanding games and their functions. Klastrup (2008) argues that in order to understand the experience of the game world we need to look at the relationship between design, meaning-making and culturalization, and to do so we need to analyze specific and prominent game world features. She does this by studying death in World of Warcraft-how it is designed as well as players' experiences of death in the game-in a phenomenological attempt at mapping game world experiences. We here argue that Livingstone's (2002) and Klastrup's (2008) approaches complement each other. In other words, to study social gameplay we can combine what happens in the actual gaming moment-what players do in the designed space-and analyze this in conjunction with their experiences and interpretations of these events. We can also take into account the community and the game culture. This would in practice imply not only studying a specific phenomenon and its related design in isolation, but also looking at actual gameplay, players, and the player community's view and interpretation of the phenomenon in order to fully understand the social side of gameplay.


2.1 Theory of social life

Social gameplay consists of players sharing a world and gaming together. Goffman defines social interaction-termed an encounter-as a situation where individuals address themselves to one another and where the encounter is reciprocally acknowledged by all participants (Goffman 1971). Each situation is governed by social norm structures which are highly contextual and culture specific. Social gameplay can be studied as encounters where social contact is essential. Gameplay can moreover be seen as focused gatherings structured by a 'sanctioned orderliness' (Goffman 1961, p.19) with local rules and identities governing the expectations and interaction of the participants. A focused gathering is an event with a certain purpose; the participants are gathered to play a game and the interaction is therefore structured around this purpose. In contrast to more permanent groups like guilds or clans, which are social arenas even when players are not actually playing together and often comprise friends or family members, the focus of temporary groups in games is on specific tasks and performing these joint, often co-located tasks is the reason for a group's existence.

An encounter between two or more people often starts with a greeting and ends with a farewell. Greetings and farewells are what Goffman calls supportive rituals. "Taken together, greetings and farewells provide ritual brackets around a space of joint activity" (Goffman 1971, p.79). Greetings and farewells respectively increase and decrease access to other individuals. When we meet someone the greeting comes first and marks a period of heightened access to the other person(s). As Goffman states, "Access, after all, is one of the things that personal relationships are about. An introduction, like a greeting, is an access ceremony" (Goffman 1971, p.79). Greetings vary in intensity and appropriateness depending on the situation in which they occur; greetings that are appropriate in one situation may not be appropriate in another. The intensity of a greeting promises something about the outcome of a situation and is defined by the expectations of its participants.

Social interaction in an encounter is not uniform but can take several forms which give the interaction different meanings. Simmel defines as forms of interaction: exchange, conflict, domination, and sociability (Simmel 1907, 1908a, 1908b, 1949).[2] Exchange is the most common form of interindividual life. Individuals exchange thoughts, love, goods, and so on (Simmel 1907). Conflict between individuals or groups leads to change and protects individuals from indifference-the absence of interaction when humans no longer acknowledge others as human beings, something entirely negative according to Simmel (1908a). Domination, another form of interaction, also can take many forms, which are united by the fact that the dominating person wants the condition of the other to be a product of her/his will (1908b). Sociability or pure sociability is defined by Simmel as the play form of association; that is, interaction free of meaning or purpose. Sociable talk, according to Simmel, is the only talk that is "a legitimate end in itself" (Simmel 1949, p.259).

Supportive rituals as well as ideas about focused gatherings and the different forms of social interaction therein can help us to structure the analysis of actual gameplay and to understand the nature of social interaction in MMO-games-how players game within the designed sociality.


3. Dungeons in World of Warcraft

This study focuses on a specific type of PUGs in WoW, 'Heroic Dungeons', which are only accessible at the most advanced level. At the end level players can set the difficulty on a dungeon or 'cave' to normal or heroic. The heroic setting makes the dungeon more difficult but also yields greater rewards. These heroics, as they are called, are very important features for gamers, both increasing the challenges of the game and its rewards; that is, the collection of items, reputation and other character improvements. These dungeons require players to join into groups of 5 for completion.

Any gamer reaching the highest level soon comes in contact with these dungeons and the rewards they offer. Many play these over and over to improve themselves and to master the challenges. In the second expansion, 'Wrath of the Lich King' (2008-11-13), there were 16 different dungeons of varying difficulty. In the third expansion, 'Cataclysm' (2010-12-07), there are 11 released dungeons so far (Patch 4.1.0, 2011-05-02). A heroic dungeon can be actively chosen and played once per day. However, there is no limit to how many dungeons can be played in one day if the player allows the game to decide which dungeon to play using the 'random dungeon option', which will be explained further below.

Looking for group (LFG) is when MMO-gamers search for others to game with. In the WoW world Azeroth, gamers LFG to find party members to do dungeons with. When WoW launched in 2004 there was only a LFG chat channel available to assist players in finding others to group with. This channel was local, meaning that only gamers in the same area of the game world could be reached. The designers later introduced a queuing system by creating so-called 'meeting stones' situated outside the dungeons which gamers had to travel to and click on to join a queue for a group. These were rarely used, however so in patch 1.9.0 (2006-01-03) the LFG channel was made global but restricted to major cities. With the first WoW expansion, 'The Burning Crusade', released in the first quarter of 2007, a new LFG system-a special menu facilitating group building-was introduced. The meeting stones were also transformed into 'summoning stones', making it possible for two players to teleport the rest of the group to their location. The LFG channel was removed and replaced with the LFG menu, but this resulted in players using the trade channel to find groups. Due to pressure from the gamers articulated on forums Blizzard later reintroduced the LFG channel but with access only through the LFG menu. In the second expansion, 'Wrath of the Lich King', a new option of choosing roles was introduced in the LFG menu. The standard roles, gamers take in a group are: Tank, Healer, or DpS (Damage per Second unit), each having a distinct play style. The highly armored Tanks are the ones engaging with enemies. Healers heal the damage enemies do to friendly players. A DpS does damage to (kill) enemies. In patch 3.3.0 (2009-12-08) the LFG tool was changed and renamed 'Dungeon Finder' (DF).

WoW is played on a multitude of different servers, where each is a separate, but identical game world with a couple of thousand players on each server. Before the introduction of the DF tool, gamers could only group with others from the same realm, making each realm a separate world where different norms and cultures could arise. Gamers had to create their own groups, often by advertising in a chat channel and then travelling to the dungeon. The new DF tool introduced cross server (cross realm) automatic grouping according to role selection, enabling players to queue for dungeons across realms and teleporting them to the dungeon, providing faster and easier grouping, with special rewards for using the random dungeon option, that is, letting the game choose which dungeon you end up in. No changes were made to this tool upon the introduction of the third expansion of WoW, 'Cataclysm'. Blizzard first introduced cross realm Player versus Player (PvP) battles with the aim of shortening waiting time, making PvP easier in low populated or unbalanced realms in terms of faction (Cheung 2006). Gamers are currently divided into battle groups, where each battle group contains gamers from several realms. Via the Dungeon Finder players can access a menu for joining a group for a dungeon with gamers in their own battle group.


4. Methodology and data

Berger and Luckmann (1991) have argued that we must move in society in order to understand it. This becomes apparent in empirical studies of MMOs, as there is little to observe unless the researcher enters into the game. In order to answer the research questions several complementary methods are necessary; it is just not possible to take all levels into account using one single method. To capture the actual gameplay an interaction analysis was chosen; and to further capture gamers' experiences and community views, interviews and content analysis of forums was performed. By employing different methods and comparing and complementing our results we are better able to understand the object of our research. Denzin (1978) suggests that there are several ways of triangulating in social research, where triangulation of methods, theories, researchers and data all are possible. Collecting data with different methods can be a strategy not only of nuancing results but also can function for method and data triangulation (Berg 2009) as a validation strategy in our research (Creswell 2007).


4.1 Interaction data

Interaction analysis has been described as "an interdisciplinary method for the empirical investigation of the interaction of human beings with each other and with objects in their environment" (Jordan and Hendersson 1995, p.39). When performing interaction analysis, video technology is often used to capture interaction so the data can be analyzed after the interaction has occurred. Fortunately, online gaming is easy to record using a variety of programs available online.

Dungeons in WoW are limited to 5 players; there is no room for outsiders to quietly observe what is happening. To study PUGs, then, our only option was to be involved in the game. At start, 24 random dungeons were played during the Wrath of the Lich King expansion. The playing sessions were distributed over weekdays (12) and weekends (12) and divided over the two fractions Horde and Alliance (6 weekends + 6 weekdays), making a total of 24 dungeons played on European English-speaking servers. A more detailed presentation of the method for this initial data gathering has been presented in (Eklund and Johansson 2010). To further complement these data, 6 dungeons were played after the Cataclysm expansion (2 horde/alliance weekend, 1 horde/alliance weekday), totalling 30 dungeons recorded.

All random PUGs were filmed and transcribed. All chat (including 'emotes' describing what the gamer is doing, e.g. Legolas laughs, on 'say' channel and 'party' channel) was recorded using a WoW add-on called WoWScribe (all chat was text based) and the data were analyzed with the organizational aid of a program for qualitative analysis (NVivo). At the final count, a total of 136 players were grouped with (not counting the researcher's characters), the number of characters being more than it should be due to leaving/replacement during the dungeons. The researcher used DpS and healer roles; a tank character was not chosen since this role often controls the pace of the dungeon run and therefore was considered to be too demanding and leading a role for the researcher. The researcher participated in all interaction but did not provoke situations. Conversation necessary for the progress of the game (e.g. asking for time to regenerate resources) was not avoided, being a part of what is expected of a gamer and in line with the participant observation method (Bryman 1995; Patel and Davidson 1994). To protect the gamers, no actual character names are used in the article.


4.2 Forum data

The forum data were collected from Blizzard's online WoW forum and processed by content analysis (Berg 2009) with the aid of a software program (NVivo). The data contain three main threads; a poll and discussion about whether the players 'have tried the dungeon finder tool yet', a thread: 'the dungeon finder tool needs to go', and last a poll with following discussion asking players what they do during a dungeon. These threads were chosen based on relevance to the research focus.

The data were analysed using the concepts of sociability and instrumentality, these tools developed specifically on the basis of the interaction data. No forum names are used in quotes.


4.3 Interview data

We performed one focus group interview with six Swedish WoW players. The theme for the semi-structured interview was social play in WoW. Three women and three men participated in the focus group, all part of the same guild and recruited through posters at Stockholm University. Some had played WoW since the release in 2004, the others for at least some years. The focus group informants were 23 to 36 years old. The interview was recorded and transcribed verbatim. The same analytical categories used for the content analysis (in NVivo) were used to perform a thematic qualitative analysis on the interview data. The names of the focus group informants have been removed to protect their anonymity.


4.4 Ethics

One major question taken into account in research involving observations, but that is problematic in a game world context, is participant consent. This problem is hard to overcome since this type of observation involves time constraints and a balance, as mentioned by Dewalt and Dewalt (2002), between observing and participating. The dungeons were completed at a fast pace and the researcher has little or no chance of asking participants for their consent before the groups was disbanded. Since consent was difficult to obtain under these circumstances, one consideration that guided us was balancing the potential risk/harm against the potential social benefit of completing the study (Berg 2009). The potential risk/harm is also considered in relation to the partial anonymity of a study where characters are recorded and not actual gamers. Here the potential risk/harm must be considered limited compared with what we gain in knowledge, as, moreover, all character names in the interaction data are anonymized.


5. Results

5.1 Interaction analysis

PuGs are focused gatherings relying on local rules and identities as well as a mutual acceptance of the ongoing situation, which builds on players fulfilling their roles and 'playing their part', otherwise the game encounter could fail. It is therefore in the interest of the group to make sure that these local rules and identities are followed (Goffman 1961). Below, DpS1 makes a mistake that could have jeopardized the encounter, but acknowledges the mistake and apologizes; showing that she is aware of the rules governing the situation.

DpS1: sorry about the pull [making the enemies attack the group], didn't cancel cast in time

Tank: no worries[3]

The results of the participant observation study show that social interaction in PUGs can be divided into two main types, instrumental and sociable interaction. Instrumental interaction deals with game strategy, e.g. asking for breaks or calling for attention when unforeseen events occur as in the example above. This interaction has the form of exchange or sometimes conflict, the common denominator is that something is at stake, e.g., the focused gathering.[4] Sociable interaction, on the other hand, deals with greetings and goodbyes, jokes, out of game discussion and discussion concerning the game but not necessary for the progression of the session. Dividing social interaction in this way allows us to understand how played sociality functions in these specific game sessions. To repeat, sociability in this context is a form of social interaction free of meaning or purpose, the play form of sociality. Some sort of interaction can be seen to take place even when there is no social interaction. For example, moving and fighting together is something experienced players can do without verbal communication (Bennerstedt and Ivarsson 2010). For it to be social interaction however, the gamers must acknowledge the existence or presence of other gamers.

An early assumption by the researchers was that there would be norms against leaving during a PUG. However, we saw a large number of gamers leaving just at the start or in the middle of a dungeon, often just after the group died during an encounter.

0:00: Dungeon starts, the tank leaves straight away. The researcher having been randomly selected as Group Leader gets a pre-programmed message saying: 'A player has left your Dungeon group. Would you like to find another to finish The Occulus?' The researcher presses "Yes" and we join a queue for a new tank. No one says anything, one player sits down and after two minutes another sits down.

4:10: A tank joins but leaves straight away; we join the queue again without words, one player, clearly bored, jumps frantically around.

4:48: A new tank joins and says 'yay' unclear why, no one replies and we start killing enemies.

There were never any reasons given for leaving and no one ever commented on it. There were no voiced social norms regulating such behaviour. This may be due to the structure of the group, with gamers from different servers who are not part of the same game world and in all likelihood will never meet again; a rule of conduct or norm would have no effect outside the actual dungeon encounter. The system that used to take norms outside the encounter-the individual reputation system in other games (Taylor 2006)-does not work here. There is no risk of obtaining a reputation as a 'quitter' or that a gamer will not group with you again, threats that have been shown to have effect in other games or even in earlier studies of WoW (Chen 2008).

10:10: After we kill the first boss [large enemy] the tank leaves directly without saying anything.

A DpS clicks 'need' on an item which is good for her. The healer also clicks 'need' although the item is no good for a healer.

The healer wins the item and leaves the party.

DpS1: freaking moron healer

DpS1: ninja.

DpS1 leaves.

Ninja is a term commonly used to describe people's in-game behavior and refers to persons who take items they do not need or that others have the right to. In this example the healer and the tank leave without explaining why and DpS1 also leaves after it is clear that there is no one left to engage with over the conflict, the encounter has broken down. Most of the recorded play sessions are lacking in any type of social interaction, gamers rarely use greetings or goodbyes and in some instances a whole dungeon could be played without a word being uttered or any non-verbal social interaction. Gamers were not investing in the social situation and as long as a dungeon ran without trouble players ignored the existence of others, making these game sessions far from social.

Social interaction was rare even in the event of downtime, refuting the notion that downtime automatically leads to socialization, as seen in the above quote where the time spent waiting for a new tank did offer the chance of sociable interaction. Nevertheless, cases with a high level of sociable and/or instrumental interaction were found, e.g. when several gamers from the same guild played together in a group. The social standard for these dungeons is set at the start, greetings acknowledge the fact that access is given to the participants and that this is a social situation as well as a focused gathering, as in the example below. During the encounter, talking continued, as seen in the example below with gamers joking about the death of a character.

DpS2: lol

Heal: hehe

In quiet contemplation, you mourn the death of DpS2

DpS2: can I wear mail [chainmail] on mage [magician]?

Researcher: ;D

Heal: sotty [sorry] cant save you from that :P

DpS2: and where is my pet?

PUG interaction is structured around the focus of the encounter, the common goals and definition of the situation. There was never any debate over this, showing that gamers share expectations on these encounters. By dividing PUG interaction into two categories-instrumental and sociable-our goal was to differentiate between interaction that supports the group strategically or else sociably. The results reveal a low level of both instrumental and sociable interaction between PUG gamers. Communication was held to a minimum and dungeons completed at a fast pace; greetings and goodbyes were by no means a certainty. In most dungeons these supportive rituals were not exchanged or only exchanged between some, showing that gamers do not invest in the social aspect of these encounters. As a final comment, we could find no difference in interaction patterns after the Cataclysm expansion of the game. The established interaction patterns remained intact.


5.2 Interviews

The results from the focus group interview show that the informants generally regard PUGs as poor social encounters. The focus group participants also expressed dissatisfaction with the instrumental play, which they felt has deteriorated because of a lack of sociable play. They did find, however, that it was easier to create groups with the help of the Dungeon Finder (DF) as it saved time. Yet they also pointed out that this was at the expense of sociable interaction in dungeons. All the informants set the DF in a context where the game itself has undergone an evolution from more social to less so and they frame the social development of World of Warcraft in terms of past and present, where the past is associated with more sociability and the present with a less stable social world. The general opinion was that WoW has moved towards including a broader audience who may not have the same opportunity to spend large amounts of time playing. This development has led to an unstable social climate. The informants argued that the less pleasant atmosphere dominating PUGs is due to a lack of norms governing social behaviour; and this lack of pleasant experiences makes gamers reluctant to invest in these social situations.

I1: You learn to not give a damn about what they say. You don't respond anymore, you just 'oh well'

I2: You go in and just do what you should and then ignore them [other players]

The informants stated that previously when gamers misbehaved, social pressure was put on gamers' guild leaders and these players were often kicked out of their guilds. In other words, there was greater social control of players by other players. The decline of these social structures has had negative consequences for the game itself. The DF is seen as a tool which enhances these processes in the game, a component further deteriorating the in-game social structures.

I1: They make it easier for all players, extend the game for the everyday player. Because of this the social part of the game disappears (I3: or is restricted at least) Yes, restricts it a lot.

Time was seen as an important aspect of this change, in the short game sequences that now are the rule when playing a PUG, gamers are not able to build relationships or establish new friendships.

I3: In the past you could keep on playing, like you said, an hour or two, or something like that just to get through it [the dungeon]. With the current situation its fifteen minutes and then it's done. So then people think 'nah, why should I get to know you, in fifteen minutes the group will split up.' There is no reason. If you keep playing for a couple of hours, three or four hours with a group, then you have to talk to people, because otherwise it will be boring. *Laughs*

The game is now easier to play on your own, since the level of reliance on other players has been decreased in later patches and expansions to make it easier for gamers to play the game without well functioning social networks, also opening up the game for more people. However, these short time spans do not lend themselves well to building relationships, whether temporary or more permanent. In the interaction data, gamers often urged Tanks on when they perceived the pace to be slow, clearly reflecting the attitude that there was no point in spending extra time with the group. In the Cataclysm dungeons, all game material that could be avoided was skipped to make the runs as fast as possible.

The gamers in the focus group argued that the decline of sociability has had a negative impact on instrumental aspects of play. Due to weak investment in the social situation, instrumental play was suffering; there is no longer patience for retries or room for failure. As could be seen in the video data, gamers often leave a PUG in case of a character/group death. Instead of engaging in a possible conflict, players withdraw from the encounter, suggesting that these encounters, devoid of social acknowledgement, in the absence of social interaction lean towards indifference. The relative ease with which these groups fall apart is one reason why the informants say that they would rather play with friends than across realms using the DF.

I4: You can be called just about anything for even a small failure. I don't like this random-group thing.

Performing demanding tasks and completing more challenging dungeons is difficult when you only have one try at a task. Furthermore, playing with people on different servers is also considered a negative factor from both an instrumental and a sociable viewpoint. The informants maintained that this makes it hard to connect with new friends that you subsequently can play again with.

I3: Then maybe before, if you found a tank that was nice and good then you added him to your friends list (number of voices agree), so the next time you just whispered to him 'Hey, are you able to play?' But now you don't do that anymore because now it's just...

I5: No, now the tank is on a completely different server. (Number of voices agree).

Getting groups together for collaborative play requires a functioning social system and opportunities to build social networks. Even purely instrumental play has earlier been shown to rely heavily on social networks (Taylor 2003). Here we see how closely linked the different types of social interaction are, in the informants' experience, and how they mutually affect each other. Sociable and instrumental interaction demand both investment and acknowledgement of the social situation. As gamers focus on completing dungeons as fast as possible even the pure instrumentality that the dungeon tool should promote suffers.

According to the gamers interviewed, the interaction-or lack of social interaction-we see in the PUGs is a symptom of a more general trend in the game. The informants also referred to Gear Score, a tool built by the players themselves, an add-on which quantifies how 'good' or 'bad' a player is based on their gear. This was seen as another symptom of the trend towards less available space for played sociality in WoW. The Gear Score system is now implemented in the game by Blizzard in the form of 'item level', governing something which used to be determined by gamer interaction. Here we see how designed sociality and played sociality interact and how the community can influence the design of the game. Although these results cannot be generalized to a broader game situation, they do provide interesting insights into the results gained from the interaction data.


5.3 Forum

The forum data interlink in many cases with the interview data; gamers express opinions on the impoverished social climate and on the gameplay benefits of the dungeon tool. The latter is more present in this material, however.

In the thread 'When running dungeons using Dungeon Finder', where the perceived effect of the DF is discussed, the gamer explains what she/he does when using it:

Forum user1:'I say 'Hello!', quietly sigh to myself as no one replies - or, if they're from Crushridge [a server], replies in Italian - and miss the days when dungeons built communities and created friendships. Then I go quiet and resign myself to simply doing my job collecting those emblems [rewards from dungeons], longing for days long since passed.'

These feelings were repeated frequently in the thread. The different servers and the structure of the communication system make it impossible to befriend other gamers and also undermine the player-constructed social norm system. There is no accountability because of the cross server groupings; there is no punishment of transgressions or fear of reputation loss.

On the other hand, gamers argued that they were indeed social, even if only when someone else initiated contact. Many stated that they just 'did their job' and avoided talk during dungeons. A minority group claimed that they liked to 'stir' things up. For example, by jumping about excessively trying to provoke reactions from other gamers. These are gamers disappointed with the social climate who try out interesting strategies to initiate interaction with other players when talking fails.

In the thread 'Did you try out the Dungeon Finder to get a group yet?, a majority of the answers praised the new tool and its functionality. The consensus was that it is much easier to get into a group. Players argued that the DF removed a boring part of the game, finding groups. Instead, gamers could focus on fun aspects like obtaining items.

There are several posts that bridge a positive attitude towards the functionality of the groups and a socially negative view. In one post, a DpS praised the DF because it made it much easier to do dungeons, while at the same time pointing out that the social situation of the dungeon was unpleasant. This DpS used to maintain an extensive list of friends to play with, but now only play with unknown gamers. The quote below from the thread 'The dungeon finder has to go' expresses the same view:

Forum user2: Dungeon finder should stay.

I, like many others, am from a low-populations realm.

There's not much activity and it's hard to get a group unless you're in a dedicated PvE [player-versus-environment] guild.

I do agree that it takes the MMO spirit out of the game

For many gamers, the functional gameplay benefits of the tool tend to outweigh the decline in social interaction.

There was much criticism in the posts concerning rules implemented by Blizzard, e.g. regarding reward distribution, which is strictly regulated and where gamers have little control. Another example is the 'kick' mechanism that removes gamers from the PUG which players experience as too difficult to use, and at the same time it is abused by players to remove participants they do not know before the distribution of items at the end of a dungeon. A third argument coincides with the interviews; because the groups are so easy to get into, the patience for multiple tries has vanished and loyalty towards the group is very low.

The general consensus was that after the cross realm groupings, PUGs were quieter and people talked less; whether or not this was considered a bad thing varied. Social gamers argued that they were affected by the mood in the dungeons, stating that they tried to talk but almost always fell silent due to lack of response. Others argued that these PUGs were not supposed to be a place for social interaction, you were there to do a job and that was all, you should simply do it as fast as you were able.

The more socially oriented gamers used the term: 'going against the spirit of the game', to signify their outlook on what an MMO should be about. There are thus two opposing views; those who focus on the multiplayer and community aspects are in conflict with those of a more functional gameplay orientation. An example of this last view from the thread 'The dungeon finder has to go' is the following:

Forum user3: We PLAY a GAME and NO, waiting for hours on end before a monitor is NOT the equivalent of 'fun'. It is frustrating, a waste of time and does not mean anything. Life is TOO SHORT.

This gamer further argued that the only community in the game is guild community. However, many gamers stressed the point that it is in PUGs that you can make friends and be invited to join guilds. This is especially true for the more 'casual' guilds (more hardcore guilds often have application forms on websites), so it would seem that casual players would be the most affected by a decline of community spirit. However, the players who promoted this view were also the ones who argued against the changes that make the game easier to play in relation to time spent.

In the threads, a majority of players claimed that the social atmosphere disappeared with the arrival of the DF. As in the interview data, a view is clearly present that the sociable aspects of the game were more 'social' before the introduction of the DF tool. Another aspect not previously mentioned is that many players perceive DF as reducing the game world, making it seem smaller. In other words, the DF tool affects how they perceive the world. Gamers do not have to travel or venture outside the populated main cities, a change which also happened in EQ and that Taylor (2006) called "from Provincialism to Cosmopolitanism". Because of this, some gamers argued that the world felt smaller and emptier. Some players went even further and argued that the DF with its instant teleport destroyed the world in World of Warcraft. Interestingly, however, these comments did not refer to socialising aspects; rather they expressed a perception of the 'instant teleport' functionality as running against the spirit of the game. This is noted by Bartle (2004), who reasons that different types of transportation heavily influence such aspects as worldliness and size of the world. In the latest expansion, Blizzard changed it so that gamers have to visit all dungeons at least once to be able to use the DF. This seems to be a diplomatic change, however, to try and please all gamers rather than something actually addressing this issue.


6. Discussion

In view of the changes to the designed structure of PUGs we conclude that investment in the social situation is low in these encounters. Goffman (1971) states that there is always a cost associated with interpersonal relationships. Such costs, concerning time, money and effort, are always involved when people decide to meet. Individuals in relationships with others develop an understanding of these costs, especially in relation to frequency and probability of encounter. Will we meet again? How much time can I spare for this relationship? A relationship where the costs are deemed high or the probability of ever meeting again deemed low will therefore be too costly for the individual and reduce the input and effort invested in that relationship. The low investment in the social situation of PUGs can be explained by this cost evaluation. PUG players come from a multitude of different independent game worlds and the design makes repeated interaction impossible. Further, gamers cannot create lasting relationships or add people they PUG with to their 'friends list'. The short time spans of these dungeons are another important factor. The cost of interaction thus exceeds any potential benefit and therefore decreases the incentives for social play. As an informant phrased it:

I6: (...) People are not as social because they can't be bothered to sit and talk. Because, you will only play for half an hour or 20 minutes and then you won't see each other anymore, so why sit and talk?

As gamers seem to learn the required strategies fast, little instrumental interaction is needed and players are assumed to know what is required of them. This produces repetition and so grinding dulls players and creates a space for indifference-non-sociality. In contrast, a study on the game Left 4 Dead 2 (Valve Corporation 2009) (albeit a very different game belonging in the first-person shooter genre) shows that randomization of the game-implemented on each course-makes pre-planned strategies useless and demands constant communication among players to succeed (Haselton 2011). The randomness makes sure players don't enter the monotone grind and keeps the game interesting as well as protecting against indifference.

Contrary to our expectations, social norms and sanctions had little impact on the social interaction in PUGs. Few instances of either positive or negative sanctions were observed. Norms were negotiated in relation to rewards but were seldom applied to other behaviour. An example of 'weak norms' was that many gamers left instantly on arrival in a dungeon. This happened on numerous occasions. No explanations were given or comments made on these leavings; as gamers leave the focused gathering breaks apart. The interview results indicate that a perceived decline in social structures makes the social situation unstable and tends to exclude adherence to social norms. Reputation and trust have been shown to be important in MMOs (Chen 2008). Reputation, however, only has a local effect on the particular server where a gamer's characters reside, since those characters are only visible to players on that particular server. The effect of ostracizing someone and reporting that character to a 'ban list' or using other sanctions such as blocking that character from future cooperation has limited or no effect across servers. Our assumption is that in this context the cost of sanctioning exceeds the gain in terms of upholding social norms (Verhagen and Johansson 2009). Psychologists have found that in game situations the ability to punish is beneficial to the group; it increases the common effort. When there was no possibility of punishing fellow gamers not contributing to the group, the gameplay suffered and the group as a whole performed less well (Fehr and Gächter 2002). Lack of sanctioning opportunities has a detrimental effect on the collaboration of the groups. The local rules of the focused gathering are upheld but otherwise the norm system in place is weak, calling for further studies on norms in different online spaces.

The cost of social interaction in PUGs is high and has a price gamers seem unwilling to pay. The low levels of social interaction we observed in the PuGs are symptomatic of a design which makes the cost of social play high, thereby creating situations where we can observe indifference. In reality, indifference is extremely hard to achieve. We can almost never stay indifferent to other individuals. But in the game indifference is shattered only when conflict arises, that is, in those situations where players engage in what went wrong rather than abandoning the encounter. The interface of the game allows gamers to attain a state of indifference where only the individual's completion or attainment of items is of interest. Egoism, Simmel writes (1908b), is free from sociation, when the other is absolutely indifferent and a mere means to ends. Indifference should not be confused with instrumentality which is neither good nor bad but simply efficient in its form. In PUGs, if the other gamers are only means to an end, egoism prevails and then sociation is not present at all. A question raised by this is whether there is something in the fact that we are playing with humans, which makes a difference in itself even if we do not act like it. In the future we should investigate this further in order to understand sociation in online venues and game worlds.

Blizzard has a privileged position, not needing to recruit new gamers for their survival; it is more than enough of an achievement to keep the players they already have. If we assume that everyone playing WoW already has a functioning social network, we can assume that the structures we see here won't deteriorate the social structure within the game. But to view social structures as fixed is a faulty assumption. We live in a social reality that is dynamic, a continuous process rather than a set one with unchanging rules (Emirbayer 1997). The social reality of any online social world is likewise prone to change, perhaps even faster than other social realities; and these changes are not only those imposed by design (designed sociality), but also those caused by user interpretations (played sociality). Both fluctuate over time and social norms along with them. Reputation is one aspect of online social worlds that have changed in WoW during the last couple of years from being perhaps the most important social mechanism (Taylor 2006; Chen 2008) to now having less meaning in some contexts. The tools introduced to govern social behaviour-which used to be controlled by norms-such as the vote-to-kick tool, item level and more, are symptoms of a design allowing for very little player agency vis-à-vis design. These are actions that players previously organized themselves, but that now have been integrated into the game by design.

As a final observation, however, these results do not communicate the qualities of the dungeon finder tool. The practical aspects of the tool empower players from low populated realms or players with weak social networks, allowing them to access further game content, and many players enjoy the functionality associated with the tool. Blizzard has recently argued that they want to make players talk to each other more and their plan was to make future dungeons more challenging, enforcing collaboration and exchange between players (Street 2011). There has been an ongoing debate over whether current MMOs, and gaming in general, is now only catering to casual gamers, as reflected in the interviews. It has been argued that increased opportunities for single play is beneficial to people with less time to play, something often connected with a more casual game style (Juul 2010). Decreasing the amount of time that must be invested in a play session is another factor that makes it easier for more players to take part in the game, opening this genre to more potential players (ibid.). However, this has clearly had a detrimental effect on players' agency towards the game structure, in some ways limiting social aspects of the game, one of the most salient features of an MMO.


7. Conclusions

Studying social interaction in temporary collaboration groups shows how important design is for user behaviour online. Changing tools and structures can both limit and expand opportunities for interaction. We show how intimately designed sociality and played sociality are connected, shaping an online social world and therefore users' experiences within this social context. In our case study, the design of temporary collaboration groups reduces the opportunity for social interaction and thus limits the possibilities for player sociality. Gamers may instead seek to collaborate only with existing friends; this could hinder reproduction or expansion of social networks. Much the same as other MMOs (Taylor 2006), World of Warcraft is progressing towards decreasing gamers' own responsibility for the social world in an attempt at facilitating for gamers with in-game tools. This is something of a balancing act since a designed tool intended to increase users' affordances may instead have an adverse effect. Studying these temporary groups allows us to observe when collaboration breaks apart and so understand when it works in digitally mediated spaces.

To conclude, it is clear in our case study that success of collaboration is suffering due to a weak social system. Gamers cannot control the social situations but are left in the hand of design tools which cannot offer the same nuanced control or ruling as user controlled norm structures and communities. A balance is needed between played and designed sociality. How users reason about these structures and how social worlds develops offer important insights into online collaboration between strangers and the implications of mediated interaction. These are important areas to study; both from the aspect of design of digital social spaces and for our understanding of online social interaction in general.



Many thanks to Jonas Linderoth and participants at Nordic DiGRA 2010 for valuable feedback and discussions.



Bartle, R. (2004) Designing virtual worlds. Indianapolis: New Riders Publishing

Bennerstedt, U. and Ivarsson, J. (2010) Knowing the Way. Managing Epistemic Topologies in Virtual Game Worlds. Computer Supported Cooperative Work, 19, pp.201-230

Berg, B.L. (2009) Qualitative Research Methods for the Social Sciences. 7th ed. Boston: Pearson Education.

Berger, P.L. and Luckmann, T. (1991) The Social Construction of Reality: A treatise in the sociology of knowledge. London: Penguin Books Ltd.

Bryman, A. (1995) Quantity and Quality in Social Research. London: Routledge.

Chen, M.G. (2008) Communication, Coordination and Camaraderie in World of Warcraft. Games and Culture, Vol. 4 (1), pp.47-73

Cheung, H. (2006-07-18) Blizzard tests new World of Warcraft cross-realm battlegrounds. Available at: http://www.tgdaily.com/games-and-entertainment-features/27605-blizzard-tests-new-world-of-warcraft-cross-realm-battleground [Accessed 5 May 2010]

Cole, H. and Griffiths, M.D. (2007) Social Interactions in Massively Multiplayer Online Role-Playing Gamers. CyberPsychology and Behavior, Vol. 10 (4), pp.575-583

Creswell, J.W. (2007) Qualitative inquiry and research design: choosing among five approaches. London: Sage Publications

Denzin, N.K. (1978) The Research act: a theoretical introduction to sociological methods. Second Edition. New York: McGraw-Hill Book Company

Dewalt, K.M and Dewalt, B. (2002) Participant observation. A guide for fieldworkers. Lanham: Altamira press

Ducheneaut, N. and Moore, R.J. (2004) The Social Side of Gaming: A Study of Interaction Patterns in a Massive Multiplayer Online Game. Chicago: CSCW '04: Proceedings of the 2004 ACM conference on Computer supported cooperative work. New York, U.S. Available at: http://dl.acm.org/citation.cfm?id=1031667 [Accessed 28 October 2011]

Ducheneaut, N., Moore, R.J. and Nickell, E. (2004) Designing for sociability in massively multiplayer games: an examination of the "third places" of SWG. In J.H. Smith and M. Sicart (Eds.), Proceedings of the Other Players conference. Copenhagen: Denmark. Available at: http://www2.parc.com/csl/members/nicolas/documents/OP-Third_places.pdf [Accessed 28 October 2011]

Ducheneaut, N., Yee, N., Nickell, E. and Moore, R.J. (2006) 'Alone Together?' Exploring the Social Dynamics of Massively Multiplayer Online Games. CHI 2006 Proceedings Games and Performances. Montréal, Québec. Available at: http://dl.acm.org/citation.cfm?id=1124834 [Accessed 28 October 2011]

Eklund, L. and Johansson, M. (2010) Social Play? A study of social interaction in temporary group formation (PUG) in World of Warcraft. Proceedings of: DiGRA Nordic 2010: Experiencing Games: Games, Play, and Players. Stockholm: Sweden. Available at: http://www.digra.org/dl/db/10343.55072.pdf [Accessed 26 July 2012]

Emirbayer, M. (1997) 'Manifesto for a Relational Sociology'. AJS, Vol. 103 (2), pp.281-31

Fehr, E. and Gächter, S. (2002) Altruistic punishment in humans. Nature 4215, pp.137-140

Goffman, E. (1961) Encounters: two studies in the sociology of interaction. Indianapolis: Ind.

Goffman, E. (1971) Interaction Rituals - Essays on face to face behavior. New York: Doubleday anchor

Haselton. M. (2011) Architecture as teambuilding in left 4 dead 2. Well Played 1 (1): 23-31

Jakobsson, M. and Taylor, T.L. (2003) The Sopranos Meets EverQuest. Social Networking in Massively Multiplayer Online Games. Proceedings of: The 2003 Digital Arts and Culture (DAC) Conference, Melbourne, Australia 2003, pp.81-90

Jordan, B. and Henderson, A. (1995) Interaction Analysis: Foundations and Practice. The Journal of the Learning Sciences, Vol. 4 (1), pp.39-103

Juul, J. (2010) A Casual revolution: Reinventing Video Games and Their Players. London: The MIT Press

Klastrup, L. (2008) What Makes World of Wacraft a World? A Note on Death and Dying. In Corneliussen H. G. and Rettberg J. W. (Eds.) Digital Culture, Play, and Identity: A World of Warcraft® READER. Cambridge, Massachusetts: The MIT Press. pp.143-166

Kolo, C. and Bauer, T. (2004) Living a Virtual Life: Social Dynamics of Online Gaming. Game Studies, Vol. 4(1)

Livingstone, S. (2002) Young People and New Media: Childhood and the Changing Media Environment. London: Sage

Myers, D. (2005) /hide: The aesthetics of group and solo play. Proceedings of: DiGRA 2005 Conference: Changing Views - Worlds in Play. Vancouver. Available at: http://www.digra.org/dl/db/06276.04321.pdf [Accessed 28 October 2011]

Patel, R. and Davidson, B. (1994) Forskningsmetodikens grunder. Att planera, genomföra och rapportera en undersökning. (Research methodology. To plan, implement and report on an investigation) Lund: Studentlitteratur

Simmel, G. (1907) Exchange. In Levine, N.D. (Ed.) Georg Simmel on Individuality and social forms. Selected writings. Chicago: The University of Chicago Press

Simmel, G. (1908a) Conflict. In Levine, N.D. (Ed.) Georg Simmel on Individuality and social forms. Selected writings. Chicago: The University of Chicago Press

Simmel, G. (1908b) Domination. In Levine, N.D. (Ed.) Georg Simmel on Individuality and social forms. Selected writings. Chicago: The University of Chicago Press

Simmel, G. (1949) The Sociology of Sociability. The American Journal of Sociology, 55(3), pp. 254-261. Translated by; Hughes, E.C.

Simon, B., Boudreau, K. and Silverman, M. (2009) Two Players: Biography and 'Played Sociality' in EverQuest. Game studies, Vol. 9(1)

Steinkhuler, C. A. and Williams, D. (2006) Where Everybody Knows Your (Screen) Name: Online Games as "Third Places". Journal of Computer-Mediated Communication, Vol. 11 (4), pp. 885-909

Street, G. (Ghostcrawler) (2011-01-13) WoW, Dungeons are Hard! Blizzard developers blog. Available at: http://eu.battle.net/wow/en/blog/1690800#blog [Accessed 10 Fenuary 2011]

Taylor, T.L. (2003) Power gamers just want to have fun? Instrumental play in a MMO. Proceedings of DiGRA 2003: Level Up Conference Proceedings. Stockholm. Available at: http://www.digra.org/dl/db/05163.32071.pdf [Accessed 28 October 2011]

Taylor, T.L. (2006) Play between Worlds: Exploring Online Gaming Culture. Cambridge: The MIT Press

Verhagen, H., and Johansson, M. (2009) Demystifying Guilds: MMORPG-playing and norms. Proceedings of DiGRA 2009: Breaking New Ground: Innovation in Games, Play, Practice and Theory. London, U.K. Available at: http://www.digra.org/dl/db/09291.50401.pdf [Accessed 28 October 2011]

Williams, D. (2006) Why game studies now? Gamers don't bowl alone, Games and Culture, Vol. 1 (13)

Williams, D., Ducheneaut , N., Xiong, L., Zhang, Y., Yee, N. and Nickell, E. (2006) From Tree House to Barracks. The Social Life of Guilds in World of Warcraft. Games and Culture, Vol. 1 (4)

Yee, N. (2009) Befriending Ogres and Wood-Elves: Relationship formation and The social Architecture of Norrath. Game Studies, Vol. 8 (2)


Games Cited

Blizzard Entertainment (2004) World of Warcraft. Blizzard Entertainment: California (PC)

Sony Online Entertainment (1999) EverQuest. Sony Online Entertainment: California (PC)

Valve Corporation (2009) Left4Dead 2. Valve Corporation: Washington (PC, XBOX 360)



[1] The first and second places are home and work.

[2] Simmel also defines prostitution as an additional form of social interaction, it will however not be dealt with here.

[3] All interview quotes are translated by the authors.

[4] We have chosen to use the term instrumental interaction rather than exchange and conflict, partly to align with earlier research and partly to analytically group interaction focused on the progression of the group task.