Of course, students do not learn how to do science only over extended periods of time through highly integrated units of study. Some topics can be treated more discretely and students can make measurable gains in a few days of instruction and practice. An example is the Klahr and Chen (2003) report on a classroom-based experiment that tested instructional approaches to teaching a control-of-variables strategy. In a short instructional sequence, students investigated balls rolling down a ramp to determine the factors that influence the distance the balls will roll. Instruction began with an “exploration and assessment” phase, in which the children were asked to make comparisons to determine how different variables affected the distance the balls rolled after leaving the ramp. Students used a wooden ramp that allowed them to manipulate two variables: the pitch of the ramp (high or low) and the texture of its surface (rough or smooth). In this phase, the children gained a base level of understanding of the phenomena and the test apparatus and were given an opportunity to think about the problem. In this study the performance of students who received instruction far surpassed that of those who did not, and their gains were sustained over time and transferred to new problems.
Whether instruction aims at narrowly defined outcomes (as the Klahr and Chen study did) or long-term investigations and a range of integrated learning goals (as did Metz’s study), there is broad agreement that children need a base level of knowledge about a domain in order to work in meaningful ways on scientific problems. Although the aims of the studies described here varied they both suggest that students need familiarity and interest in the scientific problems and that their learning requires explicit guidance. These interventions also underscore that children will often need clear statements about basic conceptual knowledge in order to succeed in conducting investigations and in learning science generally. These statements may originate from teachers “telling,” or from children reading texts, or hearing from other experts. While these intervention studies suggest that students can learn science across the strands through highly scaffolded and carefully structured experiences designing and conducting investigations, we also note that having students design and conduct investigations may be particularly difficult and require a very high level of teacher knowledge and skill in order for students to master content across the strands (see, e.g., Roth, 2002).
We elaborate on the features of problems that students investigate and the support they need to succeed in the next section. The emerging evidence suggests that learning how to design, set up, and carry out experiments and other kinds of scientific investigations can help students understand key scientific concepts, provide a context for understanding why science needs empirical evidence, and how tests can distinguish between explanations.