Skip to main content Skip to navigation

Statistics Seminar 9: Measurements, Meaningfulness and Scale Types

Files Included in this Resource:

Video recording of the speakers presentation

Further Details:

Further talks and resources in the Statistics Seminar Series can be found in the Resource Bank, or visit the Statistics Seminar Series webpages.

The seminar presentation files accompanying this talk:



Resource Description:

This session discussed the following. Data comes in all shapes and sizes. Sometimes they are numerical measurements from some electronic device and 'accurate' to several decimal places; sometimes they are counts made by a human subject to possible miscounting; sometimes they are responses to questionnaires with a '1' representing 'strongly agree' through to a '5' or '7' representing 'strongly disagree'; sometimes they are just plain text'and there are a whole host of possibilities beyond that. Recognising this is important because how one develops a statistical model and what ways to analyse the data meaningfully depend crucially on the form of the data. For instance, the mean of 2 'strongly agrees' responses, 1 'neither agree/disagree' responses and 3 'moderately disagrees' responses is clearly meaningless, yet we can do that calculation once the terms are translated into numbers.

The talks explored this, covering such ideas as nominal, ordinal, interval and ratio scales, parametric versus nonparametric statistics, and meaningful statistical modelling.


Source/Funding: The seminar introduction was given by Andrew Mead (Teaching Fellow in Life Sciences), followed by the main talk by Simon French (Director of RISCU)

Date: 26th September 2013

Additional References:

Stevens, Scale Types
Krants, Luce, Suppes and Tversky: Theory of Measurements Vols 1,2 and 3. (very mathematical)
F.S. Roberts, 'Measurement Theory', Encyclopedia of Mathematics, Vol. 7, Cambrisge University Press, 1979.

More Resources in this Series

Previous- Seminar 8: Survey Design & Analysis

Next- Seminar 10: Big Data & Large-scale Data Analysis

We would like to hear any comments about this resource, including top tips of how you have used the material to support your research development. Please post your comments below: