Measurement Issues in Quantitative Research

Abstract

Collecting quantitative data necessarily entails making decisions about how various phenomena of interest to the researcher will be defined and quantified. This is the domain of measurement and this chapter reviews several important considerations, including operational definition of constructs (and their links to theory), construct validity and measurement scales. In any behavioural research application where observations must be quantified through a process of operational definition, there is a hierarchy of four scales of measurement or quantification from which to choose. A scale merely refers to a set of numbers which can potentially be assigned to objects, people, or events to reflect something about a characteristic which they possess. Each scale possesses certain properties that determine the relationships between the numbers along the scale. In terms of increasing complexity, rigor, and degree of precision, the measurement scales reviewed in the chapter are nominal (e.g. categories), ordinal (e.g. rankings), interval (e.g. ratings) and ratio (scales having a defined zero point). The chapter also discusses how transformations from one scale to another might be done.