2008 Volume 6 Issue 1 Pages 57-63
The present study investigates whether the selection of test cases changes during hypothesis testing. In total, 75 participants solved one of two discovery tasks: the 2-4-6 task or the New Eleusis task. The participants were first presented with an initial instance produced by a particular rule and were asked to determine the rule. On each trial attempt, after the participants had formulated a hypothesis about the rule and proposed a new instance, they were provided with feedback about whether or not the instance fit the rule. The task would end when the participants announced that they were absolute confident in their best guess. The results indicate that the selection of alternative tests decreased in the latter phases, although the selection of negative and positive tests increased with time. In particular, successful problem solvers were much more likely to produce negative tests. These findings suggest that the process of hypothesis testing consists of at least two different phases.