69传媒

When it comes to teaching science, animation seems like a natural fit. What better way to convey how a bicycle pump works or the functions of the human respiratory system?

Yet research has turned up relatively little evidence that students learn more from a moving image on a computer screen than they would from a still picture or a physical model.

鈥淲e now have the hardware and software to build really beautiful instructional programs,鈥� said Richard E. Mayer, a psychology professor at the University of California, Santa Barbara. 鈥淏ut we don鈥檛 have the research on how to build educationally effective programs.鈥�

Still, an increasing number of researchers have been giving it their best shot. Their work is yielding important clues on new ways to use animation and simulation to deepen students鈥� understanding of scientific phenomena.

Seeing the Unseen

A computer model offers progressively detailed views of a molecule of DNA, from its basic double-helix outline to a close-up of its color-coded nucleotides.

*Click graphic to see the full image.

SOURCE: Concord Consortium Inc.

鈥淲e have a few stunning successes and lots of failures,鈥� said Marcia C. Linn, a professor of education in mathematics, science, and technology at the University of California, Berkeley. 鈥淏ut if we didn鈥檛 have failures, we wouldn鈥檛 be learning anything.鈥�

To educators and researchers, the potential that animation holds for science teaching is enormous. Constructed properly, they say, moving illustrations can engage and motivate students, make unseen worlds visible, and impart an intuitive feel for abstract concepts.

鈥淚 think it can really change the way science and engineering is taught,鈥� said Robert F. Tinker, the president of the Concord Consortium Inc., a Concord, Mass.-based nonprofit organization that develops innovative, interactive applications of technology for science education.

One reason that efforts to harness animation鈥檚 learning potential have largely failed so far is that the images often move too fast for people to take in all the information they convey, said Barbara Tversky, a professor of psychology and education at Teachers College, Columbia University, and a Stanford University professor emerita. A sequence of pictures or diagrams, in comparison, can be studied and revisited.

Animations can also mislead learners or distract them with color, sound, or extraneous movement, Ms. Tversky said.

鈥淭hey come out really cool, and they look great, and the chemist will say, 鈥淪ee?鈥� 鈥� she said. 鈥淚鈥檒l say 鈥楴o, I don鈥檛 see. I don鈥檛 know where my eyes should be going.鈥� They don鈥檛 understand that, for a novice, it鈥檚 meaningless.鈥�

For example, if a simulation includes text, people have to split their attention between the picture and the words, according to Mr. Mayer. Yet, without text or graphic signals, learners have a hard time figuring out what is most important to glean from the images before them.

Developing Models

Mindful of potential pitfalls, experts in recent years have begun to embed simulations in carefully designed computer-based instructional models that allow students to interact with animations, stop them in time, control them, and even create them. The Technology-Enhanced Learning in Science Center, or TELS, which Ms. Linn directs at Berkeley, offers a prime example.

With funding from the National Science Foundation, Ms. Linn and her research partners have produced dozens of computer-based instructional models that employ animations and simulations to teach students about cell division, chemical reactions, thermal conductivity, electrostatic processes, and other scientific concepts.

A psychologist by training, Ms. Linn also designs the programs to reflect research from cognitive science on how people learn.

For example, in a TELS unit on chemical reactions that is geared for high school classes, students can get a microscopic view of moving molecules in greenhouse gases and test factors that influence the formation and dispersion of those gases. The program provides cues to help students articulate and test their ideas and then critique one another鈥檚 views.

A report summing up studies of 10 TELS curriculum units, which was published in August in the journal Science, suggests that Ms. Linn鈥檚 laboratory is having some success.

For the study, Ms. Linn and her research partners recruited teachers from 16 schools in five states and tested their 6th through 12th grade students at the start and close of a school year. The tests measured the students鈥� knowledge of the basic science concepts taught in those courses.

After training, the same teachers taught the same classes again the next year to a new group of students. This time, the teachers embedded at least six of the computer-based units into their lessons. Tests were given to that group, and the researchers compared the results for both cohorts.

As measured by multiple-choice tests, the TELS students and their counterparts appeared to learn just as much science over the course of the academic year, the results showed. The TELS group excelled, however, on written tests that asked them to explain scientific ideas and link them to other concepts.

On a hypothetical 100-point test, Ms. Linn figured, the average difference would translate roughly to a 30-point edge for the experimental group.

The ability to make those connections, Ms. Linn contended, is a 鈥渕uch better predictor of future performance than whether they can remember a bunch of isolated ideas.鈥� But she acknowledged that it is also hard to disentangle which factors led to the improvement. For instance, all of the units embed scientific lessons in real-life problems.

69传媒 might have to figure out the best way to address high rates of asthma in an imaginary neighborhood, for example. Or they might be asked to interpret global-warming claims or select an energy-efficient car or an effective cancer treatment.

鈥淚 think the personally relevant problems are probably even more compelling than the animations,鈥� said Ms. Linn.

Learning From Mistakes

Others documenting some success with embedded animations include World Watchers, a middle school program that combines animations with Geographical Information Systems data developed by researchers at Northwestern University in Evanston, Ill., as well as the work of the Concord group that Mr. Tinker heads.

The consortium, with money from the NSF, has developed more than 200 interactive units that employ moving images in the teaching of science, some of which are also incorporated in TELS and other cutting-edge research projects. Educators can download the models for free at the consortium鈥檚 Web site, .

At the University of Michigan, in Ann Arbor, researchers are testing a hand-held computer tool that allows middle school students to view and create their own animations of chemical reactions.

Initial studies suggested that the program, called Chemation, was no more effective than physical models, using balls and sticks or gumdrops and toothpicks, at conveying key chemistry concepts.

But a deeper probe suggested that the lukewarm showing may have been due to a small program glitch.

鈥淲e saw kids with the animation tools model a reaction, and one end product would be a gas. There would be a single atom left over, and they would just scratch it out,鈥� said Chris Quintana, an assistant professor of learning technologies at the university. 鈥淲ith the physical model, it was just more apparent that you should end up with same number of balls that you started with.鈥�

Those studies also showed, though, that the Chemation students were more adept than a comparison group of students at recalling chemical terminology鈥攑ossibly because they had to label their creations.

鈥淎nimations have the potential to really help kids understand some challenging concepts,鈥� said Joseph S. Krajcik, a professor of science education at the University of Michigan and a collaborator on the Chemation project.

鈥淒o we have a lot of strong evidence for that? Probably not,鈥� he continued. 鈥淏ut it鈥檚 hard for me to think there wouldn鈥檛 be a learner who wouldn鈥檛 benefit.鈥�