01 July 2011

Clark Quinn on Educational Games

Engaging Learning: Designing e-Learning Simulation Games (Pfeiffer Essential Resources for Training and HR Professionals)
Clark Quinn, author of Engaging Learning: designing e-learning simulation games recently gave a masterclass in Sydney.  He introduced categories of educational games, how they differ from scenarios and simulations, and made suggestions for how they can best be used for learning.

Although the information he shared is primarily for instructional designers of elearning, rather than for interface designers, the inevitable overlap meant there was much to be gained for an elearning visual designer like me.  Here is a taste of some of his key points...

Key points from the ElNet Elearning Masterclass with Clark Quinn, 30 May in Sydney...

  • According to Quinn "Learning should be hard fun."
  • Why games? - Quinn says "games are the best practice next to mentored life practice" but that they scale better and can be less dangerous.
  • On learning styles - According to Quinn, there is no evidence that adapting learning to an individual's learning styles is beneficial, in part because we change our learning styles based on the context, and also because we don't yet have tools effective enough to make this work.  Also, there's the risk of people labelling themselves as one style or another and thus limiting themselves.
  • The complete learning experience - Quinn says games are good for a scaffolded meaningful practice but are not a complete learning experience.  They can form part of the learning ecosystem, but he emphasized the need for post-game practice and reflection to ensure that skills learned in the game transfer to the real world.  Thus there is no reason to believe that stealth learning is enough and that skills will automatically transfer outside the game environment.  Quinn suggests you spark discussions and reflection, make it personal, extend and  re-contextualise after the game. 
  • Feedback - Contextualise feedback, and keep it consistent with the worlds of the game.   Don't have the 'voice of god' come in and break the fourth wall to say "that was incorrect because..." allow clues within the game to be the feedback or change the voice of god into a response from a  game world character.
  • On failing - Challenge is essential to motivation so its important that learners fail some of the time.  (A machine learns optimally when it's failing 37% of the time.)  But Quinn warns, don't make failing too exciting (eg. lots of big explosions).
  • Virtual worlds - A virtual world is not a game by itself.  Games can be built in them.  Quinn claims they're best suited when the task is both inherently 3D and social.  American Family Insurance created a fake house in second life that has been damaged on a large scale.  Their insurance agents can explore the house and investigate the causes of damage to assess claims.  This could be turned into a game.
  • Simulation --> Scenario --> Game - A simulation is technically just a model of a system.  Put it in an initial state and ask learners to take it to a goal state and it becomes a scenario.  You can then tune that into a game.
  • Resistance to games as serious learning -  Phonics learning games for school kids were rejected by teachers in Japan who thought the kids shouldn't be playing games when they're supposed to be learning.  In the corporate world, alternate terms like "deep immersive practice" and "simulation" often substitute for "gaming" for the same reason.  Quinn noted that The eLearning guild's report on serious gaming had to be renamed "immersive learning simulations report" because they couldn't get emails past corporate spam filters if the word 'game' was included.
  • Story and narrative - Quinn emphasized that you are not telling a story in game design.  You have background story and foreground action, but there is choice.  Worry about the experience not the polish.  He also suggested designers "go wild" with that background story as learners will enjoy it more if the game world is exaggerated.
  • Assessment - A game is an assessment if they are performing in the environment and can't win until they can complete a task.

  • Sid Meier said "A game is a series of interesting decisions".  Quinn amends this to say that "An educational game is a series of important decisions"
  • Decisions v. facts - Quinn emphasises that you should think in terms of decisions, not facts or knowledge.  Ask the question "What decisions do the learners have to be able to make that they can't make now?"
  • Rules and variables - Figure out the decisions.  Capture a set of rules maybe with variable probability of different consequences built in, ideally based on known pattern mistakes and plausible wrong answers.  Look for already built models and previous research to use.
  • Settings - What are the settings in which they make these decisions?
    The more broadly you need the skill applied, the more fantastic the setting should probably be, and then, post game, ask how this would play out in a real setting or various settings.
  • Context and theme - What is meaningful to the learners? What are the experts passionate abaout? (eg. computer auditing as detective work).  
  • Characters - Henry jenkins says: choose a role that the player would like to be in. Quinn warns us to avoid stereotypes and totally flat caharacters, find small caharcter flaws that may drive their action, and write back story.  Don't show the player's character on screen, show just their viewpoint instead - show what they see. Or if you do portray them on screen, allow them to customize this avatar.  Keep in mind there's a psychological aspect to people representing themselves as they would like to be perceived rather than as they are.  See the work of Tony O'Driscoll on presence.  Avatars work best for pedagogical support, not for spouting content.
  • Game genres: Twitch (eg. racing), First and third person shooters, Strategy (eg. resource allocation), Role playing games, Massively multiplayer online rolplaying games (MMORPG).

Learning and Engagement
Quinn explains that in researching the features of learning that make it work, and the features of games that make them engaging, he identified 9 common characteristics. These include:  
  • clear goals
  • appropriate challenge
  • contextualization and thematic coherence
  • relevance (the action to the domain, and the problem to the learner)
  • available choices and possibility for exploration
  • direct manipulation
  • appropriate feedback and inter-referentiality
  • novel information and events

  • Quinn says: "If you don't get the design right it doesn't matter how you implement it. If you get the dsign right, there are lots of ways to implement it."
  • Keep it minimalist - We overproduce and overdesign.  If you get the experience right, less is needed to support it.  As in the quote by Antione de St Exupery: "A designer knows he has acheived perfection not when there is notheing left to add, but when tehre is nothig left to take away." 
  • Team - It takes a team to make a good game.  Half-bake any aspect and the level of engagement goes down (graphics, sound, dialogue, etc.) 
  • Tools - The range complexity goes from the simplest games made with Powerpoint to HTML 5 and Flash up to professional gaming on specialised consoles at the high end.  For high complexity games there is no way to do it without programmng. (eg. C and Java).  Quinn recommends Flash for most educational games.  He also recommends paper protoyping as you don't feel so bad about throwing it out and changing it.  It also allows you to user test a lot.  
  • Tuning, testing, adjusting - Will wright of Sim City says "Tuning is nine tenths of the effort".
    Programming is one tenth for a commercial level game.  For educational games, Quinn puts tuning at at least five tenths.  Play test first with yourself to spot big bugs, then someone down the hall, then your users or all types.  For learning games, you test usabiliity first. Then test the effectiveness of the learning so you can tell the difference between the causes of any problems.
  • Testing Usability -  Usability metrics include: Minimum time to complete a task,  maximum errors, how long it takes to learn the interface, how long to relearn the interface, and the subjective experience.
  • Testing Learning - Identify an independent test of what the learners are doing. After they play the game, see if they perform on the independent test.

Game examples
  • America's Army - A serious game and recruitment tool
  • Quest - built in 1995, this game to support youth at risk is still playable online 
  • CO2 FX - Brazil simulation game that puts you in the position of strategic leader trying to balance economic growth and ecological needs. In this way the consequences of the decisions made in a complex system are made visceral.
  • Virtual Leader - In this game you earn and spend credibility. Fix morale problems, finish meetings, keep people happy but all this while you move the company forward.

Other resources

Quinn's publications