The paper picks up the intensively discussed issue of potential biases on students’ evalua-tions of university teaching. Based on studies of Hamermesh/Parker (2005) and Klein/Rosar (2006) the influence of a largely disregarded factor, the attractiveness of the teaching staff on the evaluation re-sults, is in the analytical focus. As well, so far in this context extensively neglected cross-gender effects and their interaction with the teachers’ attractiveness are accounted for. In order to test these hypothe-ses deduced from a theoretical model individual data are analyzed with ordered logit multilevel regres-sion models. Despite the use of different methods, existing findings can be largely replicated. On aver-age, students especially evaluate attractive male teachers’ quality of teaching significantly better. How-ever, the results are sensitive towards different model specifications. Implications of these results for theory, methodology and higher education policy are discussed.

Keywords
Student Ratings, Attractiveness, Multilevel Models, Ordinal Variables