Measurement issues in the use of rating scale instruments in learning environment research

Abstract

The history of learning environment research is characterised by the creation and application of rating scale instruments to elicit attitudinal data from students and teachers about the learning environment. For several decades, the data from these instruments have been subject to various types of statistical analysis. Typically, such analyses are applied as part of the instrument development process as well as to inform answering of questions about the attributes of learning environments, the influences on learning environments, and the temporal stability of learning environments. Notwithstanding the widespread use of statistical techniques in quantitative learning environment research, these techniques do not necessarily reflect the knowledge and methods developed in the fields of objective measurement and rating scale analysis. It is therefore timely and appropriate to examine the implications of objective measurement theory and practice for the use of rating scales in learning environment research. The objective of this paper is to apply a measurement perspective to examine the issues involved in the development of ratings scales and in the analysis of the data they elicit. An argument is made that using raw scores from rating scale instruments for subsequent arithmetic operations and applying linear statistics is less preferable than using measures.