The conditions under which instructional animation may be effective

1. The conditions under which instructional animation may be effective Paul Ayres, Slava Kalyuga, Nadine Marcus & John SwellerUniversity of New South Wales, Sydney, Australia Paper presented at an International Workshop and Mini-conference, Open University of the Netherlands 29August 2005

2. Static diagrams v animations • Animations (dynamic representations) have not been found to be more effective than the equivalent static graphics (Hegarty, Kriz & Cate, 2003; Tversky, Morrison, & Betrancourt, 2002) • Mayer, Hegarty & Mayer (2005) found that statics were better • Koroghlanian and Klein (2004) found that more time was spent studying the information, with no learning gains • Lowe (1999, 2003) argued that learners’ attention was diverted away from key less-obvious features, suggesting a split attention type of effect.

3. Need for a theory • Some conditions, such as interactivity, improve their relative effectiveness (Bodemer, Ploetzner, Feuerlein, & Spada, 2004; Lowe, 2004; Schnotz & Rasch, 2005) Consequently, the overall results on animations are somewhat inconclusive • There is a need for a comprehensive theory to explain the cognitive processes involved in learning from animations. • With the notable exceptions of Mayer’s generative theory of multimedia learning (2001) and CLT there has been little guidance on how to design and use animations (Fred Paas). • In this presentation, I analyse animation from a CLT perspective and argue that CLT may “have some answers”

4. Cognitive Load Theory • Long-term memory (LTM) is viewed as the central, dominant structure of human cognition. • LTM incorporates a massive knowledge base • In contrast Working memory (WM) is very limited both in duration and capacity (especially when processing novel information) • However, when dealing with previously learned information these limitations disappear. Large amounts of information can be brought back from LTM to WM • The altered characteristics of WM when processing familiar information, has led to the idea of a long-term working memory (Ericsson and Kintsch, 1995)

5. 3 Types of Cognitive load(challenged at EARLI?) • Intrinsic: Element interactivity • Extraneous: Instructional interference • Germane: Investment in learning • CLT assumes that learning is inhibited if total CL is overloaded

6. Characteristics of animations • Animation has two critical characteristics that may explain its lack of effectiveness 1. Information is transitory • Unless rehearsed, material will remain in WM for no more than a few seconds • Difficult to both process current information and remember previous information • In contrast, static graphical displays can be revisited a number of times, reducing the problems associated with a limited WM • In this sense, animation produces an extraneous cognitive load

7. Retroactive inhibition 2. Animations consists of a series of successive elements • If Element A must be learned first followed by Element B, learning B may interfere with what has already been learned of A. • If Element A is to be remembered, it will need to be transferred to LTM because it will only remain in WM for a few seconds. • RI will occur if Element B must be processed before Element A has had time to be transferred to LTM • With static graphics more time can be made available to transfer the information associated with Element A to LTM prior to processing the information of Element B • Ayres & Mayer (in prep.) have found some evidence of RI with animations

8. Conditions under which animation is likely to be effective

9. Use general multimedia principles • The design of animations can in many cases be automatically improved by applying principles consistent with a generative theory of multimedia learning and CLT • Split attention etc can be avoided

10. Lowering Extraneous cognitive load • Acquiring sufficient prior knowledge • With some knowledge WM constraints caused by the transitory nature becomes less an issue • Mayer, Mathias, and Wetzell (2002) found that pre-training was beneficial in a multimedia (animated) environment • Clarke, Ayres & Sweller (2005) found pre-training on spreadsheets was effective for novices • Tracing • Counteract the transitory nature of animations and decrease extraneous load by keeping information available on the screen longer • Some evidence with Japanese writing (Marcus et al.) • Some animation types may automatically do this

11. Lowering Extraneous CL cont. • Learner control • Stop the animation. WM overload can be avoided if the learner is in control of the pace (see Mayer and Chandler, 2001) • Segmenting • If animations can be transformed in some way then they may become less demanding on WM. • Mayer and Chandler (2001) found that segmenting an animation into parts proved effective • Sequencing statics and animations • Present statics before animations for novices • However possible expertise-reversal effect for more knowledgeable learners

12. Increasing Germane Load • Interactivity • Increase interactivity which may stimulate learners to invest more mental effort in learning (see Bodemer et al., 2004; Lowe, 2004; Schnotz & Rasch, 2005) • Predicting • Hegarty et al. (2003) found that combining predictions with animations was more effective than animations alone • May not work for novices as it could increase extraneous load • May produce results similar to the use of imagination (Cooper, Tindall-Ford, Chandler, & Sweller, 2001)

13. Conclusion • Instructional animations have been introduced with little or no reference to human cognitive functioning. • As a consequence the use of computer-based animations frequently have not led to better learning environments. • To be effective new instructional technologies must be closely aligned with a knowledge of learning principles. • CLT may provide a comprehensive theory to explain why animations have frequently failed as an instructional tool • CLT may be able to identify a number of conditions under which animations can become effective.