Related studies in the field of economics have extended the definitions to encompass three types of conceptual change (Davies and Mangan 2005), namely basic concepts, discipline concepts and modeling concepts.  Many of the troublesome areas in biology, such as photosynthesis, would fall into the basic concepts category, while the larger abstract areas, such as complexity and scale, which certainly seem to incorporate discipline thresholds, would be described by the modeling concepts category.  However, working with those students already carrying out research in the field of biology may not be most appropriate to dissect out the components of a threshold, since such individuals have already joined a community of practitioners (Becher and Trower 2001), united by a common understanding or way of thinking about the discipline (McCune and Hounsell 2005).  This study therefore focused on junior undergraduate students who were first encountering concepts fundamental to the study of biology, at a level where thresholds concepts should be most evident.

A threshold concept has been described as “a keystone bringing form and robustness where previously there was a collection of ideas” (Davies 2006), and provides an interesting parallel with the concept of a keystone in biology i.e. species which have a disproportionately high influence on the functioning of living systems.  In a similar vein the concept of evolution acts as a keystone in our understanding of biological processes and interactions (Entwistle pers comm).  An understanding of evolution is clearly transformative in that it requires both a fundamental change in the way we think about living systems and an integration of knowledge from many different areas of biology. 

The aim of this study was to carry out a detailed analysis of first year undergraduate student responses to a question about evolution, to identify the defining characteristics of the threshold concept.  The question was created to provide opportunities for students to articulate a sophisticated understanding of the concept (Olsson, 1999), and a hierarchical scale of understanding was developed to score the answers, based on the qualitative SOLO (structure of the observed learning outcomes) taxonomy (Biggs and Collis, 1982).  

Of the five categories of understanding in the taxonomy, the prestructural category was clearly differentiated in responses.  Similarly, complex responses fitted the relational and extended abstract categories.  However, it proved difficult to maintain the robustness of the scale in the uni- and multi-structural categories, and a number of sub categories were created to encompass different ways of seeing and explaining the concept of evolution.  We hypothesized that this problem arose because of the existence of a threshold at this level of understanding.  A re-analysis of the responses was therefore undertaken using a phenomenographical approach, to identify educationally critical aspects of the concept (Cope and Prosser 2005). A phenomenographical approach to this investigation seemed most appropriate since this allows us to create categories of understanding in which each higher order category is inclusive of all lower order categories (Marton 1994).  Such a building of understanding mirrors the transformation and integration of ideas inherent in crossing thresholds (Meyer and Land 2003).    The analysis has produced a greater number of distinctly different categories of understanding, which take into account the essential features of evolution.  These categories give a picture of the ‘critical aspects’, as described by Cope (2006), which make up the threshold concept.  Using these critical aspects we can now design learning materials and activities which specifically show students how to make links and see relationships.  From a teaching perspective, the results of this threshold analysis show us where problems are most likely to occur in developing student understanding.  They also encourage an approach to curriculum design which makes more explicit to students the process whereby the ‘situational hooks’ integrating concepts are interlinked.