Robust Diagnostics in Regression Models for the Evaluation of University Teaching Activity

• Activity Number: 404
• Type: Contributed
• Date/Time: Thursday, August 15, 2002 : 10:30 AM to 12:20 PM
• Sponsor: Social Statistics Section*
• Author(s): Alessandra Petrucci*+ and Matilde Bini
• Affiliation(s): University of Florence and University of Florence
• Keywords: regression models ; robustness ; teaching activity

Abstract

In Italy during this recent year, there has been a great increase of the need to measure the efficiency of the educational process at the university level. As regards the research on students evaluations of the faculty teaching activity, several indicators have been proposed, and descriptive analyses as well as statistical models have been implemented. The aim of this work is twofold: 1.) to validate the questionnaire that every year the University of Florence gives to students, using appropriate indicators in regression models; 2) to detect individual students or groups of students can "affect" the fitting process of regression models and bias the results of analyses. Robust diagnostic analyses in regression models, in the sense of Atkinson and Riani approach recently proposed to regression analysis, are carried out using recent data supplied by the University of Florence.