Measurement rounding errors in an assessment model of project led Engineering Education

Abstract

This paper analyzes the rounding errors that occur in the assessment of an interdisciplinary Project-Led Education (PLE) process implemented in the Integrated Master degree on Industrial Management and Engineering (IME) at University of Minho. PLE is an innovative educational methodology which makes use of active learning, promoting higher levels of motivation and students’ autonomy. The assessment model is based on multiple evaluation components with different weights. Each component can be evaluated by several teachers involved in different Project Supporting Courses (PSC). This model can be affected by different types of errors, namely: (1) rounding errors, and (2) non-uniform criteria of rounding the grades. A rigorous analysis of the assessment model was made and the rounding errors involved on each project component were characterized and measured. This resulted in a global maximum error of 0.308 on the individual student project grade, in a 0 to 100 scale. This analysis intended to improve not only the reliability of the assessment results, but also teachers’ awareness of this problem. Recommendations are also made in order to improve the assessment model and reduce the rounding errors as much as possible.