View source for Static vs. Animated Visual Representations for Science Learning (Kaye, Small, Butcher, & Chi)

== Using Animated and Static Graphics to Scaffold Science Text Comprehension (PSLC Intern Project) ==
 ''Alyssa D. Kaye, Jenna E. Small, Kirsten R. Butcher, & Michelene T.H. Chi''

=== Summary Table ===
{| border="1" cellspacing="0" cellpadding="5" style="text-align: left;"
| '''PI''' || Kirsten R. Butcher
|-
| '''Other Contributers''' || <b>Research Programmers/Associates:</b> Alyssa D. Kaye, Jenna E. Small <br> 
<b>Co-Investigator:</b> Michelene T.H. Chi 
|-
| '''Study Start Date''' || June 2007
|-
| '''Study End Date''' || July 2007
|-
| '''Number of Students''' || <i> N </i> = 12
|-
| '''Total Participant Hours''' || 24
|-
| '''Study Type''' || Lab Study
|}
<br>

=== Abstract ===
Graphics are often used in conjunction with texts to facilitate learning, but little is known about what type of graphics are most effective for learning different science content.  This experiment sought to understand whether animated or static graphics better promote understanding and for which scientific processes animated graphics may be most beneficial.  To test this, each participant read two texts, one representing a more [[direct process|direct scientific process]] (heart and circulatory system) and one representing a more [[indirect process|indirect scientific process]] (diffusion).  Within each text, the participant viewed the corresponding animated or static graphics.  During learning, participants saw either animated graphics, static graphics, or one text with each type of graphic. Preliminary results revealed that there were no significant differences between the animated and static graphic conditions for either scientific process.

<br>

=== Glossary ===
*[[Direct process]]
*[[Indirect process]]
<br>

=== Research Questions ===
# Which type of graphic – animated or static – best facilitates [[robust learning]]? 
# Does the benefit of animation depend upon the type of scientific process - [[direct process|direct]] or [[indirect process|indirect]] -  being depicted?
# Is it better for a tutor to provide or for a student to generate motion from visual representations?

<br>

=== Background and Significance ===
The [[Coordinative Learning]] cluster defines [[coordination]] as the process of integrating relevant visual and verbal information; [[visual-verbal coordination]] is necessary to understand any type of graphic, whether static or animated.  There is strong support for a multimedia effect, in which learning from two [[multimedia sources]], or [[visual-verbal integration]], augments memory and comprehension over text alone (Mayer, 2001).  Though it is often thought that animated graphics allow for easier [[coordination]], apparent successes of animated graphics over static graphics have resulted from informational inequivalence, interactivity, and other confounds known to promote learning (Tversky, Morrison, & Betrancourt, 2002).  In addition, when learners can mentally animate static graphics, animated graphics have been found to be no more effective than the respective static graphics (Hegarty, 2004).

The effectiveness of static and animated graphics may be determined by what scientific processes they convey.  [[direct process|Direct processes]] are defined as those processes which have an identifiable causal agent and occur in a sequential, dependent manner.  [[indirect process|Indirect processes]] are defined as those processes which have no causal agent and do not proceed sequentially.  Previous research has shown that students have a better understanding of [[direct process|direct processes]] and often develop robust misconceptions of [[indirect process|indirect processes]] (Chi, in press).

Tverksky et al. (2002) have proposed that graphics are most effective when they conform to the congruence principle, which states that the structure and content of a graphic should correspond to the structure and content of the information being conveyed.  Because the circulatory system represents a [[direct process]], both the static and animated graphics conform to the congruence principle.  The process of diffusion is an [[indirect process]] involving unpredictable, random motion; therefore, the static graphics do not necessarily conform to the congruence principle and thus may be more difficult to mentally animate.

Also under investigation is the [[assistance dilemma]], which refers to the type of information a tutor should provide or withhold in order to promote optimal learning (Koedinger & Aleven, in press).  If computer [[assistance]] diminishes individual mental effort, then the student may not learn the material as completely as when the student self-generates the information.  Providing animations when students are able to mentally animate may reduce comprehension (Hegarty, 2004).  However, because [[indirect process|indirect processes]] are often poorly understood by students, it may be difficult to mentally animate these processes.  In these cases, animations provided by [[multimedia sources]] may promote optimal learning by providing information that the students would be unable to generate on their own.

This research investigates how animated and static graphics compare when other confounds, such as informational inequivalence and interactivity, are removed.  The contexts in which animations are most effective for promoting comprehension are also investigated in this experiment.

<br>

=== Independent Variables ===
*Students were randomly assigned to one of four conditions in which they read either both texts with static graphics, both texts with animated graphics, or one text with each type of graphic.  The time spent observing the animation and respective static graphics was equated so that the participant could not move to the next screen until the minimum allotted time had passed. After viewing each graphic, participants answered reflection questions designed to assess their processing of the visual.

*Variable 1: Static vs. Animated Visual Representations
**This study varied the presentation of animated or static graphics within text presented to the participant (within- and between-participants design).
**In the animated graphic condition, each participant viewed animated graphics taken from online science tutorial sites. 
**In the static graphic condition, each participant viewed six still graphics taken directly from the animations to ensure informational equivalence.
Figure 1. Example Static Graphic: Micro Diffusion 1<br><center>
[[Image:micro1_v2.jpg]]
<br></center>

Figure 2. Example Static Graphic: Micro Diffusion 3<br><center>
[[Image:micro3_v2.jpg]]
<br></center>

Figure 3. Example Static Graphic: Micro Diffusion 6<br><center>
[[Image:micro6_v2.jpg]]
<br></center>

*Variable 2: Direct vs. Indirect Processes
**Participants read two texts – one on the circulatory system ([[direct process]]) and one on diffusion ([[indirect process]]).
**A [[direct process]] is defined as one which has an identifiable causal agent and occurs in a sequential, dependent manner.
***The circulatory system represents a [[direct process]]. For example, deoxygenated blood flows from the right atrium to the right ventricle to the lungs, where it becomes oxygenated.
**An [[indirect process]] is one which has no causal agent and does not proceed sequentially.
***The process of diffusion represents an [[indirect process]]. For example in diffusion, molecules are constantly in random, unpredictable motion, which eventually creates a discernible pattern.

<br>

===Dependent Variables ===
*[[Robust learning]] was tested via a pre-test/post-test design, that included [[Retention|retention]] and [[Transfer|transfer]] items.  Students took a pretest prior to learning either test. Each posttest was administered after the learning phase for each topic (e.g. the students completed the circulatory system posttest after the circulatory system text was presented and before the diffusion text was presented).  
<br>
*''[[Retention]]'' questions tested the acquisition of information explicitly stated in the text.  This was measured by standardized gain scores from identical pre- and post-test questions as well as the percentage correct on some of the additional [[normal post-test]] questions.
**Example Retention Question: How many valves are there in the heart and where are they located?
<br>
*''Comprehension'' was defined as the level of understanding gained from the material; this was measured through various types of post-test assessments.
**''Mental model development, immediate'': Pre- and post-test participant drawings of the heart and circulatory system were used as a measure of comprehension and understanding.
<br>
Figure 4: Example of Student Pre-Test Mental Model Drawing (Single Loop 1)<br><center>
[[Image:104 pre mental model drawing_v2.jpg]]
<br></center>
Figure 5: Example of Student Post-Test Mental Model Drawing (Double Loop 2)<br><center>[[Image:104 post mental model drawing_v3.jpg]]
<br></center>
<br>
*''[[Transfer]]'' items measured deep, inferential learning. These questions tested acquisition of concepts that were not explicitly stated, and thus needed to be inferred from reading the text or viewing the graphics.  The post-test included both single-domain transfer questions and integrated transfer questions. Integrated transfer questions were presented at the end of the diffusion post-test and assessed the participant’s ability to integrate material from both texts and make inferences using that information. This deep, inferential learning was measured by percentage correct on additional post-test questions. 
**Example Single-domain [[Transfer]] Question: Some kinds of fish look for food in large groups called schools.  Each fish instinctively swims near another fish.  Swimming in schools increases the survival rate of the fish as well as the chance of finding food sources.  If the process of diffusion is similar to swimming in a school, what role does each fish play in achieving the school?
**Example Integrated [[Transfer]] Question: In the body, the concentrations of oxygen normally is much higher in the blood (outside the cells) than inside the cells and the concentration of carbon dioxide is much higher inside the cells than in the blood.  Based on your knowledge of diffusion, explain how cells receive the oxygen they need from the bloodstream and lose harmful carbon dioxide.  Be sure to mention how and why the molecules move.
<br>
*Point totals of pre- and post-tests of circulation and diffusion were equated in order to facilitate comparison.

<br>

=== Hypothesis ===
Because previous research has shown that learners are successful when they can infer motion from static graphics (Hegarty, Kriz, & Cate, 2003), we predict that static graphics will be as effective as animations when mental animation is attainable. Specifically, participants should be able to mentally animate the statics of a [[direct process]] such as the circulatory system because the static graphics conform to the congruence principle (Tversky et al., 2002).  As a result, we hypothesize that there will be no significant difference between animated and static graphics for [[direct process|direct processes]].  However, diffusion, an [[indirect process]], may be more difficult to mentally animate because the static graphics do not necessarily conform to the congruence principle.  Therefore, animated graphics should improve comprehension of the process of diffusion.  If animated graphics better support learning than static graphics for an [[indirect process]], this research would demonstrate when computer [[assistance]] might be necessary for easing the [[coordination]] of visual and verbal information.

In summary, we predict that there will be no significant difference between the animated and static graphics for the circulatory system text, but a significant difference between the animated and static graphics for the diffusion text, favoring the animated graphic condition.

<br>

===Findings===
Though there were significant learning gains for the circulatory system, as shown by the difference in pre- to post-test scores and mental model assessments (''p'' < .001), there were no significant differences between the animated and static conditions for retention or [[transfer]] assessments in this domain (all Fs < 1).  Analysis of the reflection questions also revealed no differences in the processing of animated vs. static graphics. These null findings are in accordance with our hypothesis.  For a [[direct process]] such as the circulatory system, animated graphics appear to be no more effective than mentally animating from informationally equivalent static graphics.

For the process of diffusion, there were also significant learning gains, as shown by the difference in pre- to post-test scores (''p'' < .001), but again there were no significant differences between the animated and static conditions for retention or [[transfer]] assessments (all Fs < 1).  As with the circulatory system, analysis of the reflection questions revealed no differences in the processing of animated vs. static graphics.  [[indirect process|Indirect processes]] may be less difficult to mentally animate than first predicted.

The pattern of means for the [[transfer]] scores of the diffusion assessment revealed that static graphics may be more beneficial then animated graphics, but this pattern did not reach statistical significance (''p'' = .126).   If the observed pattern continues, this may reveal that static graphics promote [[robust learning]] of the process of diffusion, demonstrating that static graphics may be more effective than animated graphics for learning [[indirect process|indirect processes]].  

All results are preliminary, as only 12 students have participated in the experiment thus far.  This experiment will continue through Fall 2007.  Future research will continue to investigate the effectiveness of animated and static graphics, specifically investigating in which domains animated or static graphics are most beneficial.  Future research may also include additional measures of [[robust learning]], such as [[long-term retention]].

<br>

===Explanation===
This experiment examines the integration of [[knowledge components]] from visual and verbal information. The experiment assesses the [[path effects]] of learning from texts with different types of graphics.  In particular, we are investigating whether animated or static graphics encourage students to employ more successful deep learning paths during the [[learning events|learning event]] and which type of graphic better promotes [[robust learning]] (as assessed by immediate [[transfer]] questions).  Furthermore, we are examining whether the effectiveness of each type of graphic differs according to the scientific process portrayed.

<br>

===Further Information===
====Connections====
<b>Visual Representations for Robust Learning in Other Domains</b>: 
*Our research seeks to better understand the effectiveness of different types of visual representations in terms of the scientific process under review.  Understanding when graphics are useful will prevent the production of unnecessary media as well as inform those who create educational materials.  This study is closely related to other PSLC projects investigating the use of visual representations to support [[robust learning]].   In [[Contiguous Representations for Robust Learning (Aleven & Butcher)]], researchers are investigating how [[visual-verbal coordination]] supports [[robust learning]] of geometry concepts.  Like the purpose of this study, Aleven & Butcher seek to inform the makers of educational technologies in domains that use visual representations, such as science or math.
*Another closely related PSLC project is [[Visual Representations in Science Learning|Visual Representations in Science Learning (Davenport, Klahr, & Koedinger)]].  In a series of ongoing studies, Davenport et al. are exploring how and why the coordination of visual and verbal information supports [[robust learning]] of chemical concepts.  
<br>

====References====
*Chi, M. T. H. (In press). Three types of conceptual change: Belief revision, mental model transformation, and categorical shift. In S. Vosniadou (Ed.), Handbook of research on conceptual change.  Hillsdale, NJ: Erlbaum.
*Hegarty, M. (2004). Dynamic visualizations and learning: getting to the difficult questions. ''Learning and Instruction, 14'', 343-351.
*Hegarty, M., Kriz, S., & Cate, C. (2003). The roles of   mental animations and external animations in understanding mechanical systems. ''Cognition and Instruction, 21''(4), 325-360.
*Koedinger, K. R. & Aleven, V. (In press). Exploring the Assistance Dilemma in Experiments with Cognitive Tutors. ''Educational Psychology Review''.
*Mayer, R. E. (2001). Multimedia Learning. New York:   Cambridge University Press.
*Tversky, B., Morrison, J. B., & Betrancourt, M. (2002). Animation: Can it facilitate? ''International Journal of Human-Computer Studies, 57'', 247-262.
<br>

====Future Plans: Fall 2007====
As only 12 students have participated in the experiment thus far, these results are preliminary.  The experiment will continue through Fall 2007.

<br>
[[Category:Study]]