Samarapungavan, cited in [#!driver-96!#], challenges Kuhn's claim that general skills of coordinating theory and evidence develop with age and are absent in early childhood. She investigated the ability of children aged 6-11 years in theory choice tasks. Children were presented with simple data about a phenomenon and given two possible theories which might explain the data. She reports that 85-90 percent of the children were able to make and explain theory choices.
Schauble et al. [#!schauble-96!#], use theoretically rich tasks for the study of scientific reasoning. Their task involved systems with fluids and immersed objects. Participants attempt to discover the causal relations between variables and outcomes. In the system, various parameters like the size, weight and cross-sectional shape of the boat, and depth of the canal, were variable. Schauble concludes that the strategies that subjects use is affected by not only by their knowledge of the skills and processes but also their conceptual knowledge base. Both these bootstrap each other.
In a recent book Theory and Evidence [#!koski!#], Barbara Koslowski legitimately argues that the prevailing view of scientific enquiry in the psychological literature carries the legacy of logical positivism. Studies of scientific reasoning are characterized by an emphasis on covariation and a corresponding neglect of theory or mechanism. She argues through a series of empirical studies that when subjects assess information, their assessments are tempered by prior knowledge about mechanism and information about alternative accounts. Most studies of scientific reasoning ignore this possibility and hence label subjects' reasoning as being flawed. Koslowski argues that the few age differences which are observed can be easily attributed to differential background information. She suggests based on her studies that students should be explicitly taught to attend to knowledge about various related areas in order to decide which alternatives are plausible and ought to be controlled in any experiment. They must also be taught explicitly not to reject a hypothesis immediately in the face of anomalous evidence. Rather students must learn to look for patterns in anomaly so as to be able to refine their working hypothesis.