26 May 2011

When to use animations for learning

picture credit: Yamaggio
You get one of those sudden urges to learn how to poach an egg or make an origami jedi master and what’s the first thing you do? I’m guessing, being the modern nerd that you are, you google for a video demo.  We’ve taken to using resources like YouTube to teach us things. Why are we drawn to learning by watching and when does it actually work?

The research into using animations vs. still pictures for learning is ongoing and there is no simple answer to the question “what’s better?”.  It depends on what’s being taught, who’s learning and what the context is.  Sound like a cop out?  Not at all, it just means we’re complex animals and you have to think a bit harder to get your answer.  But the answers are there. For example, there is very strong evidence that animations are much better at teaching human movement tasks...
Nadine Marcus of the University of New South Wales presented a seminar at the CoCo Research Centre (26 May 11) that summarized current work showing how animations are more effective for learning human movement than static pictures and why.  Here’s the gist...

Why animations are frequently less effective than static graphics

According to researchers like Marcus, Mayer, Lowe and others, the bottom line when it comes to visuals for learning is cognitive load.  Even though we’re drawn to the use of animations,  static graphics are often more effective for learning because:
  1. Animations draw our attention, but not necessarily to the right things.  Thus, they can be more likely to distract us away from the things we need to learn.
  2. Animations are transient: They change rapidly over time. The act of keeping up with something we can not control the pace of, puts a lot of pressure on our working memory.  The result: cognitive overload.
Please note, we’re not comparing an entire animation to 1 or 2 graphics.  We’re comparing an animation to an equivalent series of pictures, ideally, presented all at once.

Why do we like animations so much anyway?

Animations have a lot of visceral appeal.  Researchers think we’re drawn to them, not only for that reason, but because we are wired to learn by observation.  But what we don’t realise, is that we're wired to learn certain things by observation, not all things.  complex cognitive tasks and non-human processes are not something we learn well by observing in transient form.  When it comes to physical tasks however, that’s precisely where we want to watch someone else do it to learn it ourselves.

Why animations are better for learning human movement (motor skills)

 Studies by Marcus, Ayres, Sweller and others show that animation is actually consistently better for learning when what you’re learning is human motor skills.  (eg. manual procedures like folding origami, solving metal puzzles or bulding lego models.) (Note: I should qualify here, that it’s fine motor skills that have been tested so far. Gross motors skills are next in line.)  So why does it work so well for human movement and not for other types of learning?
  1. Human beings are wired to learn motor skills via observation.  Studies have shown that watching and thinking about doing physical things activates the same parts of the brain that are used when we actually do those things.  We can rapidly learn to copy what others do physically (which, in my view, makes a lot of evolutionary sense for a gesturing, adaptive, tool using group of primates like ourselves.)
  2. Because of this innate wiring, we can cope with more cognitive load when learning these motor skills than we can for other types of learning. It’s kind of like we’re already experts. Well, innate experts at learning motor skills through observation, just as babies are born experts at language learning. This doesn’t mean there’s no limit to what we can cope with...
Limits to our load - design and context

Animations are better for motor skills but they can still fail. Even for human movement, where animations are clearly better for learning, you can still eliminate the benefits by making the animation too long or too complex.  At this point the cognitive overload begins to take effect. The research shows that animations are only superior when they are short and simple enough for transience not to be an issue.  For human movement, we can cope with considerably more transience than we can when we’re learning other things like cognitive tasks or mechanical processes. Increasing transience (in other words overly long sequences) leads to overload. Thus, a design approach for long procedures, would be to break up the animation into manageable parts.  But only if you're designing for novices...

The Expertise factor

While information on weather patterns and science processes have been found to be harder to learn when delivered via animation vs. still pictures, this effect reverses when your learners are experts. Meteorologists are weather map experts and know what to focus on, thus they can learn from animated maps where novices would be overloaded.  Furthermore, segmentation (breaking your content into chunks) has similarly been shown to work for novices but not for experts.  Thus, you can’t make an informed decision about when and how to use animation unless you know your audience.

Design considerations

The bottom line is cognitive load. We can cope with considerably more transience when learning motor skills, and experts can manage the load better than novices, but there are still limits. So when asking the question, should this be an animation or a static graphic, or “how should I design this animation or static graphic” the underlying principle remains the same: aim to prevent cognitive overload...
  •  Make sure all the elements included are relevant, resisting extraneous decoration. 
  •  Consider the content (eg. Is it a motor skill or a cognitive task?) 
  • Consider the size of chunks (lengths of animation) do you need more segmentation or less
  • Consider the amount of time allowed for reflection/integration 
  •  Make sure the level is appropriate for the audience, knowing your design may need to be totally different for experts than for novices 
  •  Consider making the animation or video controllable and interactive to minimize the cognitive overload effects of transience (allowing the learner to watch, pause, rewatch and/or easily jump to different spots in the animation.)
Bonus track: Do hands matter?

If we’re learning to do something with our hands, is it important for us to see the hands? Marcus’ group also tested how important it is for the learning to show human hands performing a movement task. For example, if you’re demonstrating the folding procedure for an origami crane, does it matter if the paper appears to fold itself, or should the hands be there doing it? Results showed no difference between the two, as apparently we’re so adept at learning by observation we automatically infer the hands.

See also...

  • Instructional animations can be superior to statics when learning human motor skills by Anna Wong, Nadine Marcus, Paul Ayres, Lee Smith, Graham A. Cooper, Fred Paas, and John Sweller in Computers in Human Behavior, March 2009, Pages 339-347
  • Animation and learning: selective processing of information in dynamic graphics by R K Lowe in Learning and Instruction, April 2003, pages 157-17
  • The Mirror Neuron System and Observational Learning: implications for the effectiveness of dynamic visualizations by Tamara van Gog, Fred Paas, Nadine Marcus, Paul Ayres and John Sweller in Educational Psychology Review. 21-30.     
  • An expertise reversal effect of segmentation in learning from worked-out examples by Ingrid A E Spanjers, Pieter Wouters, Tamara van Gog and Jeroen J G van Merrienboer in Computers in Human Behavior, January 2011.

  • Related resources