Cluster Analysis for Mixed Methods Research
Normand Péladeau · 2021
Cluster analysis is not a statistical method per se but a vast collection of algorithms for grouping objects based on their similarity. In this chapter, the authors focus on the most commonly used clustering techniques, explain how they work, and discuss their limitations and various issues related to their use. Quantitative researchers using cluster analysis will often refer to attributes as "variables" and objects as "cases", but clustering can be used to partition either cases or variables. Although a description of the various indices proposed in the literature would be lengthy and unnecessarily tedious, the silhouette method proposed by Rousseeuw possesses specific characteristics that make it suitable for an introduction to cluster analysis. The chapter reviews some of those applications but will also stress how clustering techniques could be used differently to support additional research activities.