Priming, Consistency, Cheating, and Being a Jerk

How can developers of multiplayer games get their players to behave, cooperate, play their role, and not be such incredible jerks? I have an idea. Psychology is involved. You probably guessed this.

One of my favorite little experiments in psychology was done by John Bargh, Mark Chen, and Lara Burrows1 who were interested in how stereotypes were triggered. In one experiment, they had participants unscramble sentences that made heavy use of words like Florida, old, bingo, wrinkle, ancient and the like. A control group did the same thing, but with words not reminiscent of the elderly. That wasn’t the real experiment, though.

The important part of the experiment actually happened after the participants left the lab. Another experimenter sat in the hallway outside and discretely used a stopwatch to time how long it took participants to walk from one end of the hall to the other. Those who had been working with words related to old people actually walked significantly slower2 than those who had worked with other words.

Bargh, Chen, and Burrows also did another experiment where some people unscrambled sentences with words related to rudeness (bold, bother, brazen) and some worked with words indicating politeness (patiently, courteous, unobtrusively). All subjects then walked in on a staged scene where they had to interrupt a conversation to get some needed information. Those in the “polite” condition waited 9.3 minutes on average. Those in the “rude” condition jumped in after just 5.5 minutes on average.3

These are examples of what psychologists called “priming,” which is basically getting people in a particular state of mind or getting them to think about what you want them to. It’s a staple of advertising and surprisingly easy to do. I’ve been thinking for a while that game developers should take better advantage of it.

What if, for example, certain words of phrases were thrown around on loading screens between levels or in the matchmaking lobby for a multiplayer shooter? Would simply showing words like “sportsmanship” or “communication” or “fairness” prime people to behave themselves during games? If you didn’t want to be that transparent, you could include little stories, vignettes, or even comics or movies that included those words or illustrations of them. Or maybe you could use real data, like the number of heals provided by players in the previous game or awards for best defense. Or maybe you could include a graphic of naked, pre-pubescent angels like this:

I know you read this blog, Cliff Bleszinski. This needs to be in Gears 3. You can make it happen, man. There's still time.

In his book, Predictably Irrational,4 behavioral economist Dan Ariely suggests some even better ways of making this kind of thing work. He describes some experiments that he, Nina Mazar, and On Amir did where they asked students at MIT to solve as many math problems as they could in a fixed time. Everyone was entered into a lottery where the winner would receive $10 for each correctly solved problem, so there was incentive to answer lots of problems. Some subjects were given a chance to cheat at the task by self-reporting the number of problems solved, and some couldn’t cheat because a research assistant graded their answers.

But let’s back up a bit. Some subjects in the “cheating is possible” condition were also asked to write down the Ten Commandments before starting the math problems. The others weren’t asked to write down anything.5 Relative to those who didn’t have the opportunity to cheat, those who did but were not asked write down the Ten Commandments claimed to have answered 33% more questions –a clear indication of cheating since that’s way more than could be expected by chance alone.

But what about those who had the chance to cheat but were asked to write things like “Thou shalt not lie” and “Thou shalt not steal?” Dude, they didn’t cheat at all. They answered exactly as many questions on average as the people who didn’t even have a chance to cheat. In a follow-up study, the same researchers replicated these results by omitting the Ten Commandments and having students acknowledge understanding that their actions were “subject to the MIT honor code.” Which, ironically, was a lie; there was no such official code.

It seems that the Ten Commandments or a reference to an honor code was enough to prime people for behaving themselves, but I think the study also tapped what’s called “the consistency bias.” This is where we tend to behave in ways that are consistent with our stated intentions, especially if stated publicly.

Look, if this dude tells you to hang back and build some base defenses, you'd better listen to him.

So what does this mean for gamers? Again, I’m thinking of loading screens and between rounds of multiplayer and matchmaking lobbies. What if you presented subjects players with simple yes/no questions like these?

  • Will you change classes if your team has too many of the class you wanted to play?
  • Will you stick around to the end of the match even if it looks like you’re going to lose?
  • Are you going to curse and be rude in this next match?
  • Will you hang back and play defense if your team needs it?
  • Will you fortify your team’s defenses if needed?
  • Will you give other people a chance to drive the damn tank once in a while? Please? Pretty please? You always just drive it off a cliff, anyway. C’mon, what do you say?

If, while waiting for the match to start, each player could answer those questions, what do you think would happen? Would they be primed in good ways? Would they want to behave consistently? Would having their answers shown to other players have an effect?

Personally, I think this could work. It’s at least worth experimenting with. C’mon, someone out there try it and let us know how it goes.


1. Bargh, J., Chen, M. & Barrows, L. (1996). Automaticity of Social Behavior: Direct Effects of Trait Construct and Stereotype Activation on Action. Journal of Personality and Social Psychology, 71(2).
2. You know, like an old dude
3. …Jerks.
4. Ariely, D. (2008). Predictably Irrational: The Hidden Forces that Shape Our Decisions. New York, NY: HarperCollins.
5. Actually, there was also a condition where subjects wrote down the last 10 books they read, just to rule out the unlikely possibility that it was the act of remembering and writing something down that affected rates of cheating or math performance.

26 thoughts on “Priming, Consistency, Cheating, and Being a Jerk

  1. Alternatively, build games that reward cooperation but not cheating.

    If you look at e.g. Alterac Valley in WoW, it used to be notoriously hard to win for the Horde side. There’s a ton of reasons that could be named for that, but I really just want to focus on a few:

    – WoW requires you to grind for reputation points with various factions for good rewards; back in the day, Alterac Valley’s faction offered some of the best rewards. Getting that reputation *was* possible by winning the battelground, but it was a lot easier to do with side-quests in the battleground. Consequently, a good number of players at any time were guaranteed to not be interested in the common goal of winning the battleground.
    – Alterac Valley offers a large amount of side quests that unlock better defenses for your side; there used to be a general dispute whether unlocking those defenses works better for guaranteeing a win or simple zerging. Consequently you’re guaranteed that about half your team is not following the tactics the other half have decided on.
    – If you *have* managed to pull all your team over to the “unlock all defenses” tactics, about half your team is busy with those side quests, leaving the rest to mount a defense that’s just too meagre if the opposition decided on zerging. Dissension is bound to arise from that.

    I could go on.

    Of course there’s plain old /fun/ in seeing everyone milling around in confusion. There’s a real feeling of achievement if you pull together and /do/ pursue a common set of tactics, especially if you win.

    But I can’t help but think that instead of priming or tapping into consistency bias, it would be vastly more effective to design games with less confusing options, such that individual players are encouraged to decide that cooperation is in their best interest.

  2. I think its more likely it would work the first dozen or so times. Then people would turn their minds off and just get used to answering yes – restoring us back to the current state 🙂

  3. Halo Reach’s social matchmaking options might work for this a little bit too. I put myself down as “polite, chatty, team player” – which is generally how I behave anyway, but putting it down in an options selection should definitely help with my consistency bias.

  4. Interesting read! We’re trying some of the things you’ve mentioned in Brink as it’s a very team based game, but I’ll be the first to admit we could have gone further if doing it over again.
    For example, a small thing that I’ve been very insistent on is characters saying “Thank you” when receiving buffs. It’s been great to see how well it works in the recent consumer events.

  5. @ Neil Alphonso
    Wow, Brink dev comment ^ 🙂 I have to admit I’m intrigued by the teamwork enhancing ideas of Brink, but I’m worried that it’s all tied into carrots in the levelling mechanic – which I really hate in multiplayer competitive games for me.

  6. Pingback: Adutainment » Ist Erziehung erlaubt?

  7. One of my favorite online communities, MetaFilter, could be the site of intense flamewars, but it isn’t. I think part of the reason is that, right next to the “Submit Comment” button, there’s a little text that reads: “Remember: Everyone needs a hug.”

  8. Very good read! However, instead thinking how we could integrate this for behaviour in games, i was also asking myself about how and in which application you could use this outside of games. Like.. can farmville be perfect for a boost of collaboration at work (if you play it there in the breaks of course)?

  9. Pingback: Linkage 9 September 2010 | split/screen co-op

  10. Very interesting, I’d also like to see those kinds of sentences in games, or any other kind of effort from the devs of any game into making the player a little bit more… responsible I guess

  11. Pingback: Linkage 11 September 2010 | split/screen co-op

  12. Pingback: Simoleon Sense » Blog Archive » Weekly Wisdom Roundup 95: A Linkfest For The Smartest People On The Web

  13. I think this is a fantastic idea and one I’ve thought about before too. I hope Halo Reach’s social matchmaking options help make it a better community too.

  14. This is my example:

    I love age of conan (MMO) and I have always played healer classes in mmos. However in aoc only one stat is shown. Your kill/death ratio. You play a mini-game and nothing is shown but how many kills you get versus how many deaths. This has led many players to a mentality of “I must maximize my K/D”.

    Many will give up immediately if it seems there team is losing to preserve the k/d. In open world pvp players will run to the guards for protection as they do no want to spoil the k/d. Others simply have a guild-mate drop guild to kill them repeatedly to up their k/d ratio.

    Many have commented on the forums that if you took k/d away (or at least the deaths) it would decrease much of this behavior. Some view their k/d as an e-peen and we all know larger is better.

    I wonder if they started tracking healing done, flag caps stopped and other team oriented statistics would it help change peoples behavior in the game?

    Mind you the game is going down the toilet, this is just an example and I would not recommend the game to even my most hated enemy. Be Warned!

  15. Pingback: Culture « Where's Scylla?

  16. Pingback: Allgemein – Mit Psychologie gegen Cheater « Call of Duty Infobase – Alles über Black Ops, Modern Warfare 1 + 2, World at War!

  17. Hullo!

    I’m trying to use this info to improve my life in PUGs (especially as a healer). It’s not so much scientific studyzor (obviously) as simply trying to make LFD less of the utter abomination it is now.

    In case you’re interested, I’m recording each LFD group with a good/neutral/bad rating, and will be doing so till 31st Dec 2010.

    ^_^ Thought you might find it amusing.

  18. Pingback: Priming Your Audience To Do What You Want | Van SEO Design

  19. very good blog
    one concern of your solution (ask questions directly) is that players could get tied of answering these questions, especially again and again, so they might give up the game because of it

  20. Pingback: Theories of Instilling Moral Excellence in League of Legends - GGVogue

  21. Pingback: Leveraging the Power of Video Games

  22. Pingback: Theories of Instilling Moral Excellence in League of Legends - - Insurance, Real Estate, Banking: Loans, Credit Cards, Mortgage, Factoring, Medical, Luxury, Equipments

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.