Introduction

Science projects have played a central role in schools at least since turn of the century, when they were championed by John Dewey (1901). How can we ensure that projects are efficient, effective learning experiences that promote knowledge integration and lifelong science learning? For answers we draw on more than a decade of research by the Computer as Learning Partner project (Linn & Songer, 1991; Linn, Songer, & Eylon, 1996), and specifically on dissertation research by Clark (1996 May).

Why include science projects in the classroom? Science projects can engage students in authentic science experiences--essentially the work of experts. Projects can encourage sustained reasoning, connect classroom to personal problems, make science relevant to everyday life, and prepare students for lifelong learning. Projects give students a window into the complexities and uncertainties of science. Professional scientists engage in projects in a supportive community of mentors, peers, and skilled technicians. They benefit from shared methodologies, standards, and criteria for success. They follow sanctioned critiquing practices in reviewing each other's work and in participating in scientific meetings. How can we make these supports a part of classroom science? Our research in designing middle-school science projects in the Computer as Learning Partner project results in four recommendations.

Recommendation 1: Start with small, accessible projects
First, start small. Experts-in-training (such as graduate students) often replicate the work of others or apply established procedures before designing their own projects. In the Computer as Learning Partner eighth grade classroom, students start with projects that are slightly more complicated than the most demanding class assignment. These projects, nevertheless, require the sustained reasoning necessary to link and connect ideas, reflect on progress, and incorporate feedback.

We found three types of projects that succeed most of the time. In a critique project students evaluate an experiment or conclusion reached by another student or an account of a scientific result reported in the media. A design project engages students in building a solution to a project such as designing a house for the desert. An explain project asks students to use science principles to account for an observation such as the "dog dish" project in Figure 1, where students explain why the water in the dog dish gets warmer than the water in the swimming pool. These three types of projects allow students to generate several different ideas about the scientific question, distinguish among their ideas using evidence from class or other experiences, and draw conclusions.

Projects succeed when students can connect class experiences to their project questions (e.g., see Linn & Muilenburg, 1996). This requires the alignment of class scientific principles with both class projects and student prior knowledge. Instruction can then promote knowledge integration across all problems.

Recommendation 2: Develop class criteria for projects
Second, help students develop shared standards and criteria for the intellectual work of carrying out a project. We designed "composite" projects based on several students' work and engaged the class in critiquing them: encouraging students to identify improvements and describe weak or contradictory links among ideas or between explanations and evidence. Often students' intuitive criteria reflected superficial, schoolish standards like neatness and grammar. As a group, students developed criteria such as "back up assertions with evidence from class experiments or personal observations," and "identify confusing observations and seek additional information." These group criteria were posted and used regularly in class and individual discussions.

Recommendation 3: Provide support and coaching
Third, provide support and coaching to encourage linking and connecting of ideas and use of shared criteria. Experts carry out projects with extensive support and guidance from mentors, peers, and experts in other fields. They revise their plans based on this support. Citizens need to learn how to locate dependable coaches or experts, ways to make sense of the views of others, and strategies for incorporating useful views into revisions of their projects. To emulate this process in the science classroom, we used peer coaching, trained graduate student coaches, and teacher coaches.

We found that coaching helps students learn to monitor their progress and to make improvements to their projects. When students revise their projects, they add connections using scientific ideas, personal experience, and conclusions they have drawn. Fairly specific coaching comments, such as "what happens to light when it hits water?" or "describe a class experiment that supports your view" were more effective than general encouragement to think about related information, such as "what other variables might be influencing how warm the water gets." Students tended to add links and resolve inconsistencies in response to specific coaching comments.

In the Computer as Learning Partner classroom we synthesized the best individual coaching comments into a cybercoaching system (Clark, 1996 May). Once a project had been coached, we identified frequent student responses and we selected the most useful coaching comments for the system. The "cybercoach" allows coaches to match frequent student responses with appropriate coaching statements, and to send a message to the student. Cybercoaching took 90% less time and was almost as effective as individual coaching.
We also designed prompts based on coaching experience. These prompts raise issues, ask students to analyze their own work, and encourage them to reflect on their own progress. Using our Computer as Learning Partner software, students could access these hints or prompts for some projects. Examples of the kinds of prompts that have proven successful in our research include, "what do you need to know to carry out this project?," "what is still confusing about your results?," and "connect your conclusions to a class experiment."

Recommendation 4: Make project assessment part of learning
Fourth, make project assessment both efficient and a part of the learning process. To help students develop shared criteria for arguments and learn to critique the work of others, we structure oral project presentations. We require each student or group to prepare a project report, and to write a question in response to each project presentation. We randomly select groups to present their projects and individuals to ask questions so that each student participates at least once in the project discourse. In addition, we grade the written questions and written reports using a straightforward holistic system that rewards knowledge integration (see Figure 2).
We also assess student learning from projects through written, in-class tests of knowledge integration. In these tests, students critique, design, or explain a novel event and give the main reasons for their choice (see example in Figure 3). We score responses using the same holistic criteria found in Figure 2.

On a hot summer day Shawn's little sister notices that the water in the dogs' dish, which is sitting in the sun, feels fairly hot but the water in the swimming pool is still very cool.

If you were Shawn what would you say to your little sister to help her understand her observations?

Score Criteria

1. No science principle mentioned--descriptive only (bowl is smaller).
2. Mentions principle but inaccurate or incomplete (small things get hotter).
3. Accurately restates principle without elaboration or connections (if same heat is added than smaller object reaches higher temperature).
4. Clear and accurate understanding of single principle and adds elaboration and or context. (e.g., if same heat added to two objects then the smaller object has less space, so heat is more dense like in the lab where we heated the small and large beaker... so it reaches a higher temperature).
5. Clear and accurate understanding of principle and also ties in one or more additional principles from the same or related topic area. (e.g., the light from the sun hits the water and changes to heat energy which warms both the bowl and the pool, but since the pool has more water and surface area it doesn't reach as warm a temperature).

It is a hot summer day and Mac has invited some friends over. Mac takes two identical pitchers of lemonade out of the refrigerator and puts one on the counter in the 20 C air-conditioned kitchen and one on the picnic table outside on the covered porch where the temperature is 40 C.
.
a. Which lemonade will warm at a faster rate

(Check one)
_____The lemonade on the kitchen counter
_____The lemonade on the picnic table
_____Both lemonades will warm at the same rate

b. Fill in the blank to make a principle that applies to these pitchers:

Heat energy flows __________________________when the temperature faster / slower / at the same rate difference between an object and its environment is greater.

c. Give the main reasons for your answer.

In general, projects give students a chance to be creative in science class. Most students become engaged and carry their projects to completion, providing authentic examples of their thinking for teachers. The satisfaction of finishing a project is sufficient reward for some. Since students vary in resources for completing projects, we place more course evaluation emphasis on knowledge integration tests, oral presentations, and written questions than on the completed project.

In conclusion, classroom projects can prepare students to carry out future personally-relevant science projects,. Projects succeed when they build on what students know, starting small. Furthermore, projects are most successful when students have developed shared criteria for scientific arguments that they can apply to their own and others' work. In addition, instruction that includes coaching to stimulate reflection and revision results in more sophisticated projects. Finally, instructors can best evaluate students using projects and multiple forms of assessment. Under these circumstances, projects can engage students in sustained scientific thinking, prepare them to seek and use feedback from peers or experts, and help them systematically analyze experiments and claims they encounter in their lives.

References

Clark, H. C. (1996 May). Design of Performance Based Assessments as Contributors to Student Knowledge Integration . [Unpublished dissertation]. University of California at Berkeley, Berkeley, CA.

Dewey, J. (1901). Psychology and social practice, (Contributions to education). Chicago, IL: University of Chicago Press.

Linn, M. C., & Muilenburg, L. (1996). Creating lifelong science learners: What models form a firm foundation? Educational Researcher, 25 (5), 18-24.

Linn, M. C., & Songer, N. B. (1991). Teaching thermodynamics to middle school students: What are appropriate cognitive demands? Journal of Research in Science Teaching, 28 (10), 885-918.

Linn, M. C., Songer, N. B., & Eylon, B. S. (1996). Shifts and convergences in science learning and instruction. In R. Calfee & D. Berliner (Ed.), Handbook of educational psychology (pp. 438-490). Riverside, NJ: Macmillan.

Notes:
This material is based upon research supported by the National Science Foundation under grants MDR-8954753, MDR-9155744 and MDR-9453861. Any opinions, findings, conclusions or recommendations expressed in this publication are those of the author and do not necessarily reflect the views of the National Science Foundation.

The authors give special thanks to members of the Knowledge Integration Environment project and the Computer as Learning Partner project. We appreciate the contributions of the other group members including Steve Adams, Ben Berman, Julia Claeys, Doug Clark, Alex Cuthbert, Jeff Morrow, LaShunda Prescott, Linda Shear, Jim Slotta, Bridgette Sparks, Eunice Yi, Doug Kirkpatrick, Eileen Lewis, Nancy Songer, Jacquie Madhok, Philip Bell, Helen Clark, Betsy Davis, Brian Foley, Oliver Grillmeyer, Chris Hoadley, Sherry Hsi, Lawrence Muilenburg, Staci Richard, Jim Slotta, Erika Whitney, and Judith Stern.

Thanks also to Dawn Davidson, Liana Seneriches, Erica Peck, and Mio Sekine for assistance with the production of this manuscript.