Reflection Issue 27 | Page 22

supportive claims, examples/data, quotations and questions. To this we added the notion of epistemological comments ('epi-comments') which included numerous statements that were not about the phenomenon, but about some aspect of the learner's knowledge of the phenomenon. The main outcomes of this pilot study included a taxonomy of different types of main claims and specially epi-comments. The categories and codes were used to content-analyse the PTAs and identify trends. An example of how a patch was coded can be seen in table 2 (see appendix). One of the unexpected finding was that a third of all statements were epi-comments, about as many as main claims (see figure 1). This shows how important it is for the students to comment on the knowledge they are using. Therefore, it is possible that these types of statements help connect and make sense of their arguments. Figure 1. Statements across the four components of the PTAs for all participants. Main claims 30% 33% Supportive claims Examples Emergent questions Previous questions 1% 10% 5% 6% 15% Quotations epi-comments A second research project carried out in the second semester of 2012-13 allowed us to explore trajectories of developmental change that can take place over the 12 weeks that the students work on their PTAs. Briefly, there is an increase of epi-comments, compared to main claims, as students progress from patch 1 to the final commentary. Among the main claims, there is a tendency for initial patches to have many more claims that attempt to define the phenomenon, then in the second patch to explain it in causal terms (typical of mainstream social psychology) and later to make claims about the phenomenon from a perspective, be it a theory or model (in patch 3). In terms of the types of epi-comments, there is an increase in statements that refer to the context of ideas in pa э