Conceptual Coherence Revealed in Multi‐Modal Representations of Astronomy Knowledge

Abstract

The astronomy concepts of 345 young people were studied over a 10‐year period using a multi‐media, multi‐modal methodology in a research design where survey participants were interviewed three times and control subjects were interviewed twice. The purpose of the research was to search for evidence to clarify competing theories on conceptual coherence versus knowledge‐in‐pieces, distinguishing between coherence as revealed in the representational systems at any particular stage in a young person’s development and the changes evident in mental growth thereafter. Thus five research questions concerned with the elements and structure of understanding were investigated: (1) conceptual coherence shown as patterns of high correlation of concept representations between the media used to assess subjects’ understanding within a survey, as well as (2) coherence revealed as consistency of representation of those concepts across media and modalities; (3) enhanced conceptual understanding and skill through repeated interviews across (longitudinal) surveys, as young people develop their knowledge; (4) cultural similarity in subjects’ representations of basic static concepts (e.g. the shape of the Earth); and (5) improved understanding of basic dynamic concepts (e.g. the motion of the Earth) and complex dynamic concepts (e.g. seasons and eclipses), through “knowledge‐skill compounding”. The research findings supported conceptual coherence and rejected the counter argument of knowledge‐in‐pieces (at an alpha level of .05). Further research is recommended to replicate current research in cultures other than those of China and New Zealand studied here to confirm the view that cognition and knowledge are inherently coherent in young people.

Keywords:

• Conceptual development
• Earth science education
• Developmental research
• Physics education
• Astronomy knowledge
• Conceptual coherence
• In‐depth interviews

Notes

1. The K‐S Test is sensitive not only to differences of means and average ranks, but also to differences in the general shapes of the distributions such as dispersion and skewness (see Siegel and Castellan, 1988). If there is no significant difference between the Groups then the K‐S Test will result in a p‐value of p > .10 or p < .10. If there is a significant difference then the K‐S Test will result in a p‐value of p < .05 or p < .025, or p < .01 with increasing degree of significance (M. Coates, personal communication, July 25, 2006).

2. The results reported here are based on data of the Extended Survey Group and Extended Control Group from the first, second, and third Surveys in New Zealand and China.

3. By coding and categorisation scheme we mean a system of schema for classifying participants’ cosmological concepts (e.g. eclipses, seasons) on ordinal scales from least to most scientific to afford statistical analysis. The scheme was based on descriptors and thumb‐nail sketches encapsulating the essential features of each concept, represented through a variety of media, and readily understood by coders (see Bayerl, Lüngen, Gut, & Paul, 2003). Similar classification schemes have been used extensively in the field (see Nussbaum, 1979; Nussbaum & Novak, 1976; Sneider & Pulos, 1983; Vosniadou & Brewer, 1992, 1994).

4. Individual identities have been disguised in the names used in the following section.