«Static and animated presentations in learning dynamic mechanical systems Jean-Michel Boucheix*, Emmanuel Schneider ** Learning and Development ...»
Learning and Instruction 19 (2009) 112e127
Static and animated presentations in learning dynamic
Jean-Michel Boucheix*, Emmanuel Schneider **
Learning and Development Studies Laboratory, French National Science Research Organization, University of Burgundy,
Building AAFE, Esplanade Erasme, BP 26513, 21065 Dijon Cedex, France
Received 3 July 2007; revised 21 December 2007; accepted 3 March 2008
In two experiments, we investigated how learners comprehend the functioning of a three-pulley system from a presentation on a computer screen. In the ﬁrst experiment (N ¼ 62) we tested the effect of static vs. animated presentations on comprehension. In the second experiment (N ¼ 45), we tested the effect of user-control of an animated presentation on comprehension. In both experiments the participants were university students. Comprehension was measured with a test including three comprehension indicators. The ﬁrst experiment indicated that an animation as well as integrated sequential static frames enhanced comprehension.
The second experiment showed that a controllable animation did not have a powerful effect on comprehension, except for learners with low spatial and mechanical reasoning abilities.
Ó 2008 Published by Elsevier Ltd.
Keywords: Static pictures; Animation; User-control; Spatial ability; Comprehension; Mechanical reasoning ability; Expertise reversal effect
1. Introduction In spite of the recent explosion of animated websites in education, the cognitive beneﬁts of animated illustrations in the understanding of dynamic mechanical systems included in technical or scientiﬁc documents remain disputable ´ ´ (Betrancourt & Tversky, 2000; Tversky, Bauer-Morrison, & Betrancourt, 2002). While a small number of studies show positive effects of animation in understanding complex systems (Hidrio & Jamet, 2002; Mayer, 2001; Rieber, 1990, 1991; Rieber, Tzeng, & Tribble, 2004) other studies show little or no effect (Harrison, 1995; Kinze, Sherwood, & Loofbourrow, 1989; Lazarowitz & Huppert, 1993; Mayer, Hegarty, Mayer, & Campbell, 2005; Palmiter & Elkerton, 1993; Pane, Corbett, & John, 1996) or even a negative effect (Lowe, 1999, 2004; Schnotz & Grzondiel, 1999). This study aimed at investigating which types of static and animated illustrations presented on a screen can enhance the learning of a dynamic mechanical system.
* Corresponding author. Tel.: þ33 80 39 57 65; fax: þ33 80 39 57 67.
** Corresponding author.
E-mail addresses: email@example.com (J.-M. Boucheix), firstname.lastname@example.org (E. Schneider).
0959-4752/$ - see front matter Ó 2008 Published by Elsevier Ltd.
doi:10.1016/j.learninstruc.2008.03.004 J.-M. Boucheix, E. Schneider / Learning and Instruction 19 (2009) 112e127 113
1.1. Efﬁciency of animations One of the reasons why animations fail to improve comprehension of dynamic systems is related to their frequent violation of the ‘‘apprehension principle’’ deﬁned by Tversky et al. (2002): ‘‘the structure and content of the external ´ presentation should be readily and accurately perceived and comprehended’’ (p. 256; see also Betrancourt, 2005). An animation conveys changes in the depicted phenomenon as well as movement over time. However, perceptual and cognitive processing limitations constrain the comprehension of a changing visual situation (Ayres & Paas, 2007;
Mousavi, Low, & Sweller, 1995). Moreover, changes may regard different elements of a dynamic mechanical system while the movement of the elements is simultaneous. Furthermore, learners often have to share attentional resources ´ between text and animation. Thus, animations are often very hard to perceive and to process (Betrancourt, BauerMorrison, & Tversky, 2001) and pose strong perceptual, conceptual, and working memory demands (Ayres & Paas, ´ 2007; Betrancourt & Tversky, 2000).
One reason why learners do not beneﬁt from animations is probably the insufﬁcient ways in which spatial and temporal changes are communicated. More ‘‘apprehendable’’ external displays accompanied by text and illustrations could improve the building of accurate mental representations of dynamic technical systems. Within this framework, we conducted two experimental studies in order to investigate if two kinds of cognitive aids, based on static and animated presentations accompanied with text, improve comprehension of a dynamic mechanical system. The ﬁrst cognitive aid was based on a static presentation that conveyed the key steps of the functioning of a dynamic mechanical system. The second was based on interactive processes between the animation and the learner, and regarded the efﬁciency of user-control of the course of the animation.
1.2. Illustration of dynamic information
One possible way to produce a more ‘‘apprehendable’’ representation of a dynamic mechanical system could be to design, instead of an animation, multiple static frames which discretely, but precisely, depict the key steps of a dynamic process. In this case, the static design is employed in order to support the learner’s cognitive simulation of the dynamic process. Previous research has shown that a sequential external presentation of a dynamic process can bring more beneﬁts than animation (or as much as an animation) in enhancing mental representation (Hegarty, 1992, 2004; Hegarty, Kriz, & Cate, 2003; Mayer et al., 2005; Paas, Van Gerven, & Wouters, 2007; Zacks & Tversky, 2003).
However, other recent studies have shown that a continuous animated presentation outperforms a static presentation ´ (Betrancourt, Dillenbourg, & Clavien, in press; Catrambone & Fleming Seay, 2002; Thompson & Riding, 1990). To explain the contradictory ﬁndings one could pose the question if it is the nature of the elements of a dynamic mechanical system that makes the difference on how they are presented and, consequently, learned.
Hegarty et al. (2003), using the toilet cistern system, showed that multiple static illustrations of the main phases of the functioning of the system, which showed how the device works, improved mental representation and increased comprehension compared to single animated or single illustrations. In a study concerning three-pulley systems (Hegarty, 1992), the participants had to perform a task regarding the kinematics of the system based on a static picture.
The results indicated that to infer the movement of the lower pulley (at the end of the causal chain) the participants inspected the upper pulleys (at the beginning of the causal chain). They inferred the components’ motion ‘‘beginning by imagining the rope being pulled and working through the causal chain of events in the motion of the system’’ (Hegarty, 2004, p. 282). This piecemeal strategy seems consistent with a discrete representation of the local events.
Mayer et al. (2005) compared the effectiveness of presenting different types of content via sequences of paper-based static pictures (accompanied by text) with presenting the same content via computer-based narrated animations. No advantages were found for animations across the four content-types examined (lightning storms, toilet cistern, ocean waves, and car brakes).
However, inferring motion transitions between discrete static pictures could be difﬁcult. Thompson and Riding (1990), using a lesson about the Pythagorean Theorem, showed that the performance of participants who worked with a continuous animation was better than that of two other groups who learned the steps of geometrical transformations using a discrete multiple presentation of static graphics on paper or using a single static graphic. In learning computer algorithms, Catrambone and Fleming Seay (2002) showed that an animation was a better aid than a discrete static ´ graphic presentation taken from an animation. Betrancourt et al. (in press) found that a continuous animated presentation of lightning storms outperformed a discrete static-frame presentation in a retention and transfer test.
114 J.-M. Boucheix, E. Schneider / Learning and Instruction 19 (2009) 112e127 Participants in the continuous animated presentation spent more study time than those in the discrete static-frame presentation, thus suggesting that a series of static frames did not always lead to strong inference activity.
The contradictory results observed when comparing continuous animation and sequential static frames could be partly due to micro-step granularity and to the conditions under which the sequences of static frames were presented.
Concerning granularity, Tversky et al. (2002) noted that ‘‘many of the static graphics portray only the coarse units’’ of a process or of an object ‘‘whereas the animations portray both the coarse and the ﬁne segment’’ (p. 252). However, multiple graphics should depict more ﬁne-grained segmentation, without adding too much information, to avoid cognitive overload. The presentation of sequential static frames is a crucial point because inferential activity could depend on it.
1.3. Static vs. animated presentation
In previous research the conditions of static sequential presentation were not always speciﬁed. Static frames can be
presented according to two conditions at least. In the ﬁrst case each static frame appears independently of the others:
the ﬁrst frame disappears when the second appears, and so forth. Under such circumstances, the new information may override the old information in working memory. Such a sequential independent-static-frame presentation may also interfere with the perception of continuity of the movements and also increase cognitive demands because the learner has to maintain the ﬁrst frame in working memory while the second is processed (Ayres & Paas, 2007; Paas et al., 2007). Moreover, it can be assumed that sequential independent-static-frame presentation could interfere with inferential activity by increasing extraneous cognitive load (Paas et al., 2007).
In the second case, integrated sequential static frames are presented side by side on the same screen. The entire segmented process is available simultaneously for the learner. This presentation can support the building of a precise dynamic mental representation. Each step of the process is visually accessible on the same screen, allowing the learner to mentally create and maintain continuity in the perception of the process. Moreover, direct visual comparisons between the different steps are possible, enhancing germane load (Ayres & Paas, 2007). Such comparisons may generate inferences about steps of ﬁner granularity between the main coarse steps. It can be assumed, then, that this type of integrated sequential presentation of static frames is particularly effective for the processing of information ´ involving elaboration of the mental representation of dynamic mechanical systems (Betrancourt & Tversky, 2000;
Tversky et al., 2002).
1.4. User-control of animation
Another way to present more apprehendable animations, recently explored by a number of researchers (Boucheix, 2008; Boucheix & Guignard, 2005; Kriz & Hegarty, 2004; Lowe, 2004; Mayer & Chandler, 2001; Price & Rogers, ´ 2004; Schwan & Riempp, 2004; Tassini & Betrancourt, 2003), regards user-control on the animation’s course.
Interactive displays provide the opportunity to stop, rewind, and restart, to slow down, or to change directions.
Interactivity, from the point of view of memory demands, should lead to less cognitive load and should improve comprehension. User-control also gives the learner the opportunity to replay a part, thus shaping the display. A learner can adapt the display speed to his/her own rate of cognitive processing. A ﬁrst basic level of user-control regarding the rate of presentation of multimedia slides has been studied by Boucheix and Guignard (2005) and Mayer and Chandler (2001). These studies showed the advantage of this level of user-control.
More sophisticated procedures in user-control of animation were tested by Schwan and Riempp (2004). They tested two levels of interactivity. Speciﬁcally, students learned to tie four nautical knots of varying complexity by watching either controllable or non-controllable video presentations. In the controllable presentation, participants could stop, reverse, replay and modify the video’s speed. In the non-controllable presentation the speed and duration of the video were ﬁxed and participants could only restart the animation from the beginning. The results showed that the controllable presentation led to a strong reduction in the cognitive load related to the processing of the task.
However, allowing full control of an animation is not necessarily always effective. Novices may not be able to use the animation’s interactivity features effectively when faced with complex tasks or systems e the same regards, for example children (Boucheix, 2008). In the ﬁeld of meteorological maps, Lowe (2004) found that novices who worked with user-controlled weather maps, animation was ineffective because they tended to focus upon features that were perceptually salient rather than thematically relevant (see also Bogacz & Trafton, 2005). Consistent with these results, J.-M. Boucheix, E. Schneider / Learning and Instruction 19 (2009) 112e127 115 Kriz and Hegarty (2004, 2007) found that students with low prior relevant knowledge were unable to build an adequate model of how a novel device works using controllable animation. In studies of undergraduate students’ learning about ´ brain synapse mechanisms, Betrancourt et al. compared non-controllable, partly-controllable and fully-controllable
´ ´ ´versions of animated presentation (Betrancourt & Realini, 2005; Tassini & Betrancourt, 2003). They found that usercontrol of the animation did not produce superior comprehension; the best results were obtained with the noncontrollable version.
The above contradictory results have been found with widely differing contents: from procedures which presuppose imitation to predicting weather which presupposes a high level of