Sunday, July 26, 2009

What Broken Windows and a Fake Prison Can Teach Us About Player Behaviour


One of the most interesting messages of Malcolm Gladwell's The Tipping Point is that context plays a greater role in human behaviour that most acknowledge. As one of the central tenets of game design is anticipating and shaping player behaviour, it's important to understand the ramifications of this claim.

This is contrary to the belief most people have about personality and decision making. When observing another's behaviour, most see that person's decisions as indicative of what "kind of person" they are and discount circumstantial influences. In psychology, this is known as Fundamental Attribution Error.

But context matters a great deal. Take the Stanford Prison Experiment, where a small mock prison was setup with half the subjects as guards and the other half as prisoners. Otherwise very healthy and stable people (including some who identified themselves as "pacifists") assumed the behaviours of strict disciplinary guards and riotous prisoners with shocking celerity. An atmosphere of terror was established very quickly and due to the extreme stress most of the subjects were exhibiting, the two-week experiment was ceased after only six days.

A similar experiment inspired by the biblical parable of the Good Samaritan was conducted by two Princeton psychologists on a group of seminary students. The seminary students were given a questionaire about their motives for studying theology and were then told to prepare a quick speech on a biblical theme. Some were told to discuss the relevance of professional clergy, others were told to discuss the story of the Good Samaritan. Finally, the experimenters told the students on the way to deliver their speech that they were either slightly late and should hurry or that they were a few minutes early, but it was probably fine to go anyway.

On their way to give their extemporaneous speech in a different building, all the students passed by a man slumped against a wall, eyes closed, coughing and groaning and clearly in need of assistance. So which factors influenced the seminary students to stop and give aid? It wasn't their motives for studying theology, or even having just prepared a speech on the story of the Good Samaritan. No, the only factor that influence their behaviour was whether or not they were told they were late. Of the students told they were running late, only 10% stopped to help. Those that were told they had time to spare, 63% gave aid.

The Broken Windows theory states that where policing does not enforce minor crimes (vandalism, graffiti, fare jumping in subways, etc.) actually creates an environment where people are more likely to commit serious crimes.

If context does affect behaviour with such significance, what implications does this have for designing experiences with different types of player behaviour in mind?

First, we may need to consider that a player's Bartle's Type might not be a completely definite and discrete categorization. A "Killer" type player might say they enjoying fighting other players, but this might simply be because there are some other qualities (aesthetic, demographic, etc.) that draw them to games where behaving like a Killer is rewarded. The same player might enjoy a game that doesn't particularly provide Killer-esque gameplay, but for others reasons, they simply aren't play that game.

Rather than designing toward supporting each Bartle type, it might be the case that we need to provide the context and reward for a specific type of behaviour and then find ways to attract different types of players. Some of the most beloved games (e.g. Portal, Shadow of the Colossus, Mario) aren't successful because they support a great diversity of behaviours, but rather because their context and the behaviours they encourage are harmonious.

To reflect this back to recent discussions of creating meaningful decisions, understanding the importance of context also provides more tools for framing decisions. If we want to create a situation where resisting a violent decision is difficult, we should create an environment that makes violence seem reasonable. If a decision hinges on agreeing or disagree with a large group that holds one opinion, an environment of homogeny will likely affect this decision.

We need to be aware of the effects of context if we want to allow players to make decisions that are truly their own. Otherwise, we might be subtly driving many players toward a certain decision without even realizing it. We actually may want to do this, for a certain aesthetic or artistic message. But this decision should be deliberate, rather than as an unintended consequence of environment art, music or dialog.

As always, thinking about the effects of context isn't a mandate. It's simply another tool to have in our designer workbench. There will obviously be certain games where the effect of context is more pronounced than others. But regardless of the type of game we're making, I think it is important to remember that who we imagine the player to be might be just as important as the circumstances we create for them. And we would do well not to underestimate the potency of those circumstances.

Labels: , , ,

11 Comments:

Blogger Graham said...

I've often been curious about how the Broken Window Theory applies to games. I've participated in a lot of MMOs and online worlds and online games in general, and I've always found it startling how some of them are so rotten and have a tendency towards being jerks, and some are hugely civil, self policing, and friendly.

I've wondered what kinds of factors create these differences. Clearly, these games are doing something different. I've even noticed it in myself -- some games I want to be part of the community, some games I want to just gain points and crush or brush those in my way.

Really, there are obvious things that games do to encourage this behavior. Even more interesting, is wondering how often the developers are in control of those things, or even considering them!

July 26, 2009 at 3:45 PM  
Blogger Nels Anderson said...

@Graham I'm definitely curious about how deliberately different MMOs are building things toward certain behavioural goals. Is it really intentional, or does it just sort of arise out of the player demographics that are outwardly attracted to the game? If I had to guess, I'd say it was more of the latter.

July 27, 2009 at 7:53 AM  
Blogger Scott Juster said...

Another interesting psychological study that came to mind when reading this piece was the Milgram Experiment. While less focused on circumstantial influences on decision making than the Stanford prison experiment and the broken window theory, it demonstrates that people can be persuaded to make decisions in the context of being asked to do so by an authority figure.

While I haven't ever applied an electric shock to any of my fellow players (maybe we've discovered a new use of the Wii Vitality Sensor?), I am often amazed at how easy it is to let myself just passively accept whatever the game tells me to do. "Collect some widgets? Why not? Kill 50 orcs? You're the boss!"

I think Bioshock's real triumph was pointing out how much trust we actually place in what "the game" (i.e. "the designers") tell us to do.

Game designers have a uniquely authoritative position, which can easily create a circumstance in which the player (consciously or unconsciously) experiences a power imbalance.

July 27, 2009 at 12:03 PM  
Blogger Nels Anderson said...

@Scott Bioshock really did pull that off masterfully. And the second time I played through Far Cry 2, the very first mission you get it to go kill some guys guarding a hut. Why? Because someone told you to.

Blindly doing whatever you're told in games is at least partially due to many, many games requiring it to progress in any meaningful way.

But just about the worst thing a designer can do in this regard is requiring the player to do something they know they don't want to do just so a character can "betray" them later. If that ham-fisted technique never appeared again, I'd be pretty content.

July 28, 2009 at 6:55 AM  
Blogger Jorge Albor said...

An interesting aspect of the Milgrim Experiment is the proximity to authority figures. How affecting the command of the authority, in our case the developer or in-game character, may depend on the perceived proximity to the participant.

The can be said for how explicit they are in the roles they are to fill a la the Prison experiment. Those participants were told quite explicitly the role they would fill, and they expanded that role with their own perceptions of guards from movies and books.

Rambling. Anyway, now say a developer drops you into the wilderness with no explicit role or tool, a perfect emergent play ground. In this environment, maybe their Bartle type is more important. Unless the developer is really sneaky with their design.

July 28, 2009 at 11:51 AM  
Blogger Nels Anderson said...

@Jorge I think that's probably right. Perhaps because the participants in the Stanford Prison Experiment had some archetypal understanding of what is prison is like, they subconsciously began to adopt those roles. Games do the same, except that archetype is almost always a gruff, take-no-prisoners, one-man army.

I wonder too if many gamers have adopted a sort of "meta-archetype" where they play games like they're playing a game. I.e. we've become so familiar with some of the tropes, we behave in accordance to those tropes, rather than in response to the game itself.

I've seen this happen in tabletop RPGs, where the grognard player behaves strangely as a character, but intelligently as a player of a game. I actually find it quite refreshing playing with relatively novice games, since they're able to come into that role without any prior knowledge about how to "best" play the game.

Gah, I don't even know if that's clear. I may have to think about this more and turn it into a full post or something ;)

July 30, 2009 at 9:12 AM  
Blogger Jason T said...

I just started Far Cry 2 (mostly based on your permadeath post), and it really made me uncomfortable that the first thing I was asked to do was to kill a couple guys outside a hut. On the way there, all I could think was: What'd these guys do? Am I the kind of guy who just kills for money? I barely know this protagonist and I'm already agreeing that he's a killer. And starting The Darkness the other day, I even paused a moment to question why my protagonist didn't seem to react more strongly to having demons sprout out of him. I guess Bioshock and Shadow of the Colossus really ruined me in this way, encouraging me to question why things in games were happening at all.

I think your point about grognards on tabletop RPGs is spot-on, and I wish that newbies to gaming were used more often to test whether games make sense to those who aren't already predisposed to just go with the flow.

August 3, 2009 at 1:51 PM  
Blogger Nels Anderson said...

@Jason Not sure if you heard the latest Experience Points podcast (from Scott and Jorge upthread, incidentally), but there's a great discussion about how the success of Bioshock, or at least its legacy.

Similarly, I think Bioshock succeeds on two fronts. The setting, story told through aftermath, etc. is all fantastic. But it also precisely targets how absurd it is that we always, unthinkingly do as we're told in games.

Some may criticize Bioshock for not answering the question it poses, but honestly, no game could feasibly do both. It does raise the stakes for 2K Boston's next project though ...

In any case, I wish more gamers (myself included) could just turn off our conditioned responses and approach same games more tabula rasa. Obviously some games require a more "refined pallet," of course, but others I'd really like if people approached without all the baggage of their previous games.

August 4, 2009 at 8:51 AM  
Blogger mike said...

This comment has been removed by the author.

August 5, 2009 at 9:55 PM  
Blogger Mike Darga said...

Great post. I love to see people tying these psychology concepts into games.

The example of Broken Windows that comes to mind pretty readily is Barrens Chat =)

August 5, 2009 at 9:57 PM  
Blogger Nels Anderson said...

@Mike Haha, definitely. The context provided by a game's community can definitely feedback into how new members of that community behave. "Permission" given (or not) by peers factors pretty dramatically into how people behave in certain situations.
E.g. apparently the contrast between the communities in League of Legends and Heroes of Newerth is quite ... stark.

While this isn't objectively good or bad, it's something that's more or less unavoidable and I'd say it's unwise to ignore the ramifications.

August 6, 2009 at 12:58 PM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home