If the aim of education is for all students to learn and improve, then the expectations must be transparent to the students. In this study, three aspects of transparency are investigated in relation to an examination methodology for student teachers: self-assessment criteria, a scoring rubric, and exemplars. The examinations studied were carried out in 2004, 2005, and 2006 respectively, all with a cohort of first year student teachers (n = 170, 154, and 138). There was a large difference in scores between the 2004 and 2005 cohorts (d = 3.21), when changes in the examination were implemented. The comparison between 2005 and 2006, when no changes were made, does not show a corresponding difference (d = 0.27). These results suggest that, by making the assessment more transparent, students' performances could be greatly improved. (Contains 1 figure, 4 tables and 1 note.)