Traditionally, categories that arise from a phenomenographical analysis
are not predetermined; they emerge from data (Akerlind, 2005). This
investigation, however, applied both self-efficacy theory as well as
phenomenography. Initial, segmented coding was therefore guided by a
predetermined set of codes – Bandura’s (1997) four predicted sources of
efficacy beliefs. Phenomenographical analysis was used in the analysis of how
students’ experienced these various sources of efficacy. Analysis began by
viewing interview excerpts assigned to each type of efficacy source as a set,
rather than as individual transcripts. This was done to ensure that the focus of
the analysis remained on interpreting students’ collective experiences, an
important focus in phenomenography (Akerlind, 2005).