Shining a light on NOIR - Rethinking Scales of Measurement
Speaker: Professor Chris Brunsdon (Maynooth University)
Venue: Michael Smith Lecture Theatre, Dover St, Manchester M13 9PT
Abstract: The 'NOIR' referred to here is actually an acronym for the four scales of measurement proposed by Stevens in 1946: Nominal, Ordinal, Interval, and Ratio. Despite being proposed over 70 years ago this categorisation is still influential - and can shape the way people think about choices for data analysis and visualisation. A number of software packages and text books are structured at least loosely on this framework - many of which have informed practice in spatial modelling and mapping. However, it has not gone unchallenged and it will be argued here that this approach is at times unhelpful. There are inconsistencies arising from the NOIR categorisation, and it omits several important scales of measurent. However, I would argue that it is still helpful to consider scales of measurement, although this involves critically evaluating Stevens’ original specification. In this talk, I will discuss practical examples of these inconsistencies and omissions, and consider specific cases of analysis of some of the omited scales of measurement - in particular cyclic data and partially-ordered sets.
Bio: Chis Brunsdon is a Professor of Geocomputation, and Director of the National Centre for Geocomputation at Maynooth University. Prior to this he was Professor of Human Geography at the University of Liverpool in the UK, and before this he worked in the Universities of Leicester, Glamorgan and Newcastle. He holds degrees from Durham University (BSc Mathematics) and Newcastle University (MSc Medical Statistics, PhD in Geography).