employed, and the degree of coherence of children’s alternative ideas about vision and colour. An open-ended questionnaire and then a closed-ended questionnaire about vision and the concept of colour were administered to an intact classroom of 32 fifth-grade pupils. Six of the pupils were subsequently selected and interviewed (using the 'traditional' clinical interview) about the same topics first individually and then as a group. The same students were again interviewed individually "about an event," where a colour light source was used to change the colour of objects. Next, the six pupils were asked to complete the two initial questionnaires first in pairs and then as a group. Pupils' discussions in both cases were audiotaped. The data analysis showed that children felt hesitant to express their ideas during the individual clinical interviews, and that their ideas were very simplistic. Pupils demonstrated, however, knowledge of light-depended colour change of objects, which they never attempted to investigate, because they did not recognize these experiences as interesting. There was also evidence indicating that children’s ideas did not remain unchanged during the interviews. Peer group discussions proved to be more 'ecologically' valid, but they did not appear to provide sufficient evidence for the assessment of an individual child’s conceptions. They provided however useful information about the way individuals construct knowledge as members of a dynamic group such as the classroom. The results of the study converge with the findings of other studies, which provide support to the claims that children's alternative conceptions in science lack consistency and coherence, and that the interviewing techniques themselves may affect and distort pupils' initial understandings.
Keywords: Alternative ideas, interviewing techniques, contstructivism, colour of objects.