Public Release: 

Carnegie Mellon Researchers Say Direct Instruction, Rather Than "Discovery Learning" Is Best Way To Teach Process Skills In Science

Carnegie Mellon University

This material relates to the session "How Scientists Really Think: Beyond the Myths of Discovery"

PITTSBURGH--Direct instruction using the Control of Variables Strategy, rather than discovery learning, may be the best way to teach young children about science, says a Carnegie Mellon psychologist who is conducting a four-year field study in public schools in Pittsburgh, Pa.

The field study could lead to a new kind of science curriculum for elementary schools.

CVS is the skill that allows scientists to design unconfounded experiments and to draw valid conclusions from experimental outcomes. However, for all of its importance, CVS is not something that most children acquire naturally, even if they have a lot of exposure to discovery learning experiences. Instead, CVS is a cognitive process skill that must be taught.

Early lab testing and classroom work to teach the Control of Variables Strategy, or CVS, to young children yields promising results, says Carnegie Mellon Professor David Klahr. He explains that for children to become scientific thinkers, somewhere along the way they must obtain a set of skills for comparison and solid logic that allows them to recognize potentially important experimental results.

With post-doctoral fellow Zhe Chen and Anne Fay, a former student now teaching at Stanford University, Klahr has examined the effectiveness of different methods for teaching elementary school children the Control of Variables Strategy -- one of the central process skills needed in scientific experimentation.

"Our work demonstrates that even preschool children can distinguish between knowing an answer, when there are no confounds or alternative explanations, and having to guess when there are confounds and alternative explanations," Klahr said.

Klahr says it is important to teach children this kind of process skill with direct instruction, using concrete examples and hands-on tasks so that the underlying logic is understood and the principle can be applied to a broad variety of topics. Simply exposing children to lots of undirected experience in designing experiments won?t give them the necessary feedback to distinguish a good from a bad experiment. Klahr claims that "in teaching this is type of skill, discovery learning is usually ineffective and inefficient."

"The trick is to find ways to get children to understand that what they learn about a well-designed experiment in one area -- say biology -- has the same underlying logic as a well designed experiment in some other area, such as physics or social studies," Klahr explains.

By helping children to acquire a reasoning skill fundamental to scientific experimentation, teachers may be able to help their fledgling scientists actually design experiments and interpret their outcomes.

With a $725,000 grant from the James S. McDonnell Foundation and participation by several schools in the Pittsburgh region, Klahr and Chen are developing, implementing and assessing a set of instructional materials for teaching elementary school children the concepts and skills of scientific inquiry.

Klahr and Chen began by gathering data from children in second through fifth grade about how they analyze simple experiments such as what happens when springs are stretched or balls roll down slopes. Their goal was to teach these children how to design experiments that compare situations or conditions that differ in only one variable while keeping all other variables constant so that valid conclusions can be drawn. Then, through a series of questions to elicit children's explanations of their own experimental designs, Klahr and Chen are able not only to determine how kids think an experiment will play out but also to discover why they think a result will occur.

In some cases, they found that after children were taught basic principles for designing good tests with specific devices, they were capable of designing experiments with other devices, and making appropriate evaluations of other people's experiments. Less than 30 minutes of direct instruction was sufficient to teach most children CVS skills that were still in place seven months later.

But Klahr says teachers are rarely able to take the time to isolate a process skill like CVS from the particular substantive topics they are trying to teach, such as plant growth, mechanics or electricity. So CVS skills, even in elementary schools that have good science programs, are usually not solidly mastered by most young students.

Klahr's goal is to create a CVS curriculum that is based on solid research from cognitive psychology. Their long-range plans include developing instructional strategies that will be useful in teaching this fundamental skill in the context of a rich and varied set of science topics.

The effects of various types and levels of instruction will be examined and children's performance on various transfer task (e.g. evaluating and designing tests and explaining outcomes of tests) will be assessed.

"Together, the lab and classroom findings will form the basis for the design of new curriculum units that are grounded in cognitive theory and will be integrated into the classroom context," Klahr says.

For a related story:


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.