Robust benevolent evaluations of teaching performance

Although benevolent (BoD) weighting has some appealing features for the construction of evaluation scores of teacher performance (i.e., SET-scores), it still suffers from an important drawback in its basic form: the sensitivity of its outcomes to the influence of potential outliers, extreme values, and potential measurement error in the data. The robust order-m BoD approach provides a solution for this problem as it no longer puts central the traditional assumption that all observations (thus also potential outliers or observations infected by measurement errors) should be considered in the computation of the SETscores. As a result, teacher evaluations become less affected by outliers and incorrectness in the data. The main purpose of this paper is to present this order-m BoD approach as a robust alternative to basic BoD and discuss some of its advantages to evaluate teachers. The method is illustrated using data collected at the Hogeschool Universiteit Brussel (Belgium).