**5. Discussion**

Disagreements within the research community about the most appropriate methods of analysis of concept maps have the potential to inhibit the widespread classroom use of the tool to support learning [43]. The benefits and drawbacks of traditional quantitative or qualitative approaches to map analysis are compounded by the fact that researchers have not previously discriminated between the types of knowledge that have been embedded within maps. The application of the semantics dimension of Legitimation Code Theory o ffers a new approach that is explicit in the need to consider understanding to be composed of qualitatively di fferent knowledges that need to communicate with each other in the pursuit of expertise.

The consideration of the degree of semantic density and semantic gravity exhibited within map propositions o ffers a more nuanced consideration of map quality that is achieved by considering map morphology alone. However, it allows for the consideration of that which is 'yet-to-be-known' (rather than assessment of 'correctness') so that maps of contested values and beliefs can be assessed using the same approach as maps of agreed factual content [17]. The significance of the semantic profiles that students exhibit in their concept maps o ffers a window into some of the issues they experience on their educational journey—particularly as they move between school, university and professional life. For example, the di fferences in the structuring of knowledge that exist between a high school and a university biology curriculum (that has been observed by Kelly-Laubscher and Luckett [46]), sugges<sup>t</sup> the existence of a possible mismatch between the semantic range that students are expected to navigate at university against that which they will have experienced in secondary school. This may cause problems for students' transition from school to university when their school education is assumed to have given them the necessary prerequisite knowledge to embark upon their undergraduate studies. Tracing the changes in the semantic profiles that students exhibit provides a visualization of the progress that students are making against desired outcomes, o ffering a way of monitoring student progression and curriculum e ffectiveness. However, we cannot assume homogeneity of the knowledge quality held by students as they enter university, even when they have covered the same content at school. The two examples shown here display di fferences in students' semantic profiles such that the student represented in Figure 4 appears to exhibit a greater semantic range within his knowledge structure of this content area, suggesting a better preparedness of undergraduate study. To confirm this, we need to explore a greater range of curriculum content with the students to see how key areas of the curriculum have been structured in the students' minds.
