Learning the statistical analysis of measurement errors

Abstract

This paper describes the conceptions of first‐year University students, about ‘measuring’. A teaching sequence has been provided, involving a theoretical lecture course on the analysis of measurement errors, followed by laboratory work in optics and in electricity. Data have been gathered during the laboratory work. Their analysis shows that the students establish a hierarchy in the series of measurements and do not really understand the need to make several measurements. Few of them have critical insight into the notion of confidence interval. Random and systematic errors are not distinguished. The general view is that, the more measurements one makes, the ‘better’ the result is, without understanding the nature of this ‘better’.