An important issue in developmental research is whether children’s ideas in a given domain display conceptual coherence or whether they constitute knowledge-in-pieces (see Vosniadou, 2008 and diSessa, 2008, respectively). Our earlier work (Bryce and Blown, 2006; 2007; Blown and Bryce, 2006) compared the broadly similar developmental patterns of boys and girls in two contrasting cultures (New Zealand and China) with regard to basic astronomy knowledge. The research revealed the extent of the developmental coherence of children’s ideas looked at longitudinally. The current studies followed the same group of young people (those who were accessible, N = 345) for a further period of 5 years to investigate more advanced astronomical knowledge, including the dynamic concepts of the Earth’s seasons and eclipses (requiring subjects to bring about a mental rotation of several objects; e.g., imagining the movements of the Earth and its Moon around the Sun). The study considered both the conventional view of a concept as a mental model or representation (Gentner & Stevens, 1983; Mandler, 1998) and newer interpretations of a concept as a skill; that is, regarding a concept as a creative ability rather than as a simple recall from memory (Barsalou, 2003). The investigations involved semi-structured, one-to-one, Piagetian interviews requiring the researcher to spend considerable periods of time in each country, enabling the young people involved to accept his presence, and interpreters and local assistants in the case of China in order to translate audio transcripts and maintain the necessary links and contacts with the participants in several school districts. This research involving in-depth, multi-media interviews certainly ties in with Barsalou's idea of a concept as a skill and, with it, the notion of conceptual coherence as an ability to consistently create more or less identical concepts using different media of representation involving different sensory modalities (Blown and Bryce, in preparation).

The data collection began with the participants’ direct observations of the movement of stick shadows in sunshine and developed with (initially open) questions put to them concerning their observations, about the apparent motion of the Sun, and the appearance and motion of the Moon, leading to explanations of the motion of the Earth. The interviews had separate parts devoted to gathering verbal data; then sections where subjects drew their ideas using pencil and crayon; and finally sections where they were encouraged to represent their ideas through play-dough models. Records made of the interview protocols were subjected to close analysis involving the development of category schemes and ordinal scales into which subjects’ ideas could be placed for the concepts involved. A range of correlation and other analyses were used to check for similarities and comparisons, particularly for the detailed understandings revealed through using several media and modalities.

The statistical results showed strong evidence of conceptual coherence as patterns of high correlation of cosmological concept categories between and across media of representation. There was little difference between the two cultural groups in basic static cosmological concepts, like Earth Shape. However, in the case of dynamic cosmological concepts, there was evidence of improved understanding by the China Groups over the NZ Groups. These results can probably be explained by differences in social policy (e.g., the reported intellectual advantage of only children over children with siblings in modern China, coupled with the related high expectation of teachers and parents, and fierce competition for places in high status universities); together with curricular emphasis (e.g., the advantage of more time being allocated to astronomy) in the case of the China Groups; and experience/conceptual skill in the case of both China and NZ Survey Groups.