Performance measures for peer assessment
Rafael Molina-Carmona, Rosana Satorre-Cuerda, Patricia Compañ-Rosique, Faraón Llorens-Largo
Universidad de Alicante
Fourth International Conference on Technological Ecosystems for Enhancing Multiculturality (TEEM2016)
https://2016.teemconference.eu
Salamanca, November 2-4, 2016

Abstract:
Peer review is an evaluation strategy increasingly used in educational environments. It has now a central role in the field of teaching, but there is an extensive experience in the field of research. This evaluation method implies an author reviewing, usually anonymously, the work of his or her colleagues who, in turn, can become reviewers of the author’s work. This peer evaluation system enriches the evaluation itself, and it can even become one of the few ways to assess the works in absence of a hierarchical authority on the subject. In massive access learning platforms it has even become the main assessment method. All in all, it is necessary to assure the quality of the assessment process. In this research we consider the process of peer review as a classification process, in which we have several classifiers (reviewers) that, before an input (work to revise) should assign a particular class (work grading). In this analysis two performance measures are proposed to assess the degree of goodness of the reviewers, in order to help improve the quality of peer review processes providing a more objective assessment of the reviewers task.