An important purpose of student evaluation of teaching is to inform an educator's reflection about the strengths and weaknesses of their teaching approaches. Quantitative instruments are one way of obtaining student responses. They have traditionally taken the form of surveys in which students provide their responses to various statements using item-by-item agree/disagree ratings. Previous research has identified shortcomings of such rating scales, including response bias and the associated lack of discrimination amongst the items evaluated. In this paper, best-worst scaling is proposed as a novel method for quantitative teaching evaluation. The way in which best-worst scaling can be used in this context is illustrated in three different applications. Two applications demonstrate how it can be used for evaluations in a small-size classroom environment. The third application is a broader evaluation of university courses on a larger scale. In comparison with conventional rating scales, the best-worst scaling approach enables better highlighting of the differences between evaluation items. In doing so, it can provide enhanced guidance to educators in their reflection about their teaching. Moreover, implementation and analysis of a best-worst scaling evaluation is relatively straightforward, which establishes it a feasible method for teaching practitioners and researchers.