an elementary school in Cyprus. The 12 students were divided into two equivalent groups based on their school performance. Two different devices were used for the interviews, one for each group. Each device consisted of a wooden white box with eight coloured light bulbs lined up across the top of the box surface and a line of switches across the middle of the box surface. Four or five switches were attached in each device, respectively, and they could move in two positions, up and down. A battery, the bulbs, and the switches were connected in a "hidden" circuit inside each box. In the circuit, one of the switches was connected as a general switch, while, in the second device, the fifth switch was a dummy one. During the interviews, students were instructed to think aloud and explain their thoughts prior to conducting any experimentation in their attempts to find out the way each lamp would light on. They were also instructed to proceed in a step-by-step fashion and were encouraged to keep records of their experimental results. This procedure was based on Klahr and Dunbar’s (1988) study, who consider scientific investigation as a dual search, a search in a space of hypotheses and a search in a space of experiments. The results showed that both spaces varied depending on the number of switches on the device, and that children searched more frequently the experimental space, when trying to secure information for the solution of the problem. Children also had a tendency to disregard evidence contrary to their hypotheses, and were not always able to use the control-of-variables strategy or combinatorial reasoning. Prior experiences and the context of the problem also seemed to influence students' strategies. These results have implications for the design of appropriate teaching scenarios promoting cognitive growth and for future research.
Key words: cognitive abilities, problem solving, scientific investigation, dual space search.