Abstract
Previous research shows a discrepancy between different teachers’ assessment of student theses. This might be an even larger problem in the context of teacher education, since teacher trainers originate from different disciplines. This study aims to investigate how different assessors prioritise between criteria for assessment. Criteria were elucidated using repertory grid interviews with teacher trainers. These criteria formed the basis for a web-based Qsort that was distributed to teacher trainers at three different universities. The 66 teacher trainers in this study represented different disciplines, but were all engaged in the supervision and examination of student theses from teacher education programmes. The analysis revealed three different factors, which have been labelled: logic text structure as product, research process as product and results as product. The differences may not be explained by the background (e.g. discipline, university or gender) of the teacher trainers. Together, these three factors explain why teachers assessing student theses may accentuate different personal criteria. Hence, this might impact on their grading practice and also the interrater reliability.