Segments: Multivariate Segmentation Analysis
Content
Data analysis is usually concerned with the relationships between multiple variables across a group of cases. Such an approach is sometimes referred to as R-methodology, examples of which include regression, analysis of variance and structural equation modelling.
Some problems cannot, however, be adequately addressed in that way, but require instead a focus on the relationships between cases across a group of variables, sometimes called Q-methodology. Although the two approaches are intimately linked (they use the same data matrix) they yield different, though not incompatible, perspectives on the empirical world.
The clinic first discusses the prototypical research questions for which each approach is most suitable, how to recognise which is most appropriate to a given project, and the differences and relationships between these two forms of multivariate analysis. It then introduces a number of Q-analysis methods.
The most familiar is likely to be ‘cluster analysis’ a set of methods to discover inductively homogeneous groups of cases in a data-set. Traditional cluster analyses suffered from a number of problems. The clinic will therefore focus on more recent methods that solve some of these problems such as optimal segmentation (CHAID), and latent class analysis.
Finally, strategies for combining the R and Q analysis approaches in practical research are discussed and illustrated.
This clinic will be supplemented with online learning materials which can be accessed after the event.
Prerequisites
Introductory understanding of statistics, e.g. the mean, standard deviation and standard error of a variable. If you are attending the clinic as part of our Researcher Development Initiative then you automatically fulfil the prerequisites.