QCA vs cluster analysis

[Sarah Wang]

Hi everyone, 

Hope everyone is well and happy. Could I please ask a question related to comparing cluster analysis and QCA? 

I understand these two methods are ultimately different regarding theoretical underpinnings or the statistical operation and calculations. 

Is it okay to say "QCA also answers the question of what cases are more similar to each other (when we look at the sufficient solutions)"? (Of course, more importantly, QCA also answers what configurations are associated with the outcome).  

Could you please also suggest some reading in this area (cluster analysis and QCA comparison)? 

Thank you all

Sarah 

[Ingo Rohlfing]

Dear Sarah,

for the assessment of case similarity and dissimilarity, the truth table would be enough and, in my view, better than a sufficient solution. First, cases that share the same conditions can differ in the outcome membership, which is not directly apparent from the conventional truth table. Second, a solution is minimized truth table information and often interpreted in causal terms. This is different from establishing (dis)similarility between cases in a descriptive analysis. Using the truth table for descriptive purposes is usually referred to as a typological analysis, I think. Cases that are members of the same row are of, or belong to the same type. 

This paper compares fsQCA and fuzzy clustering:

https://doi.org/10.1080/13645579.2010.483079 There is also a chapter on QCA and cluster analysis in this handbook that may be of interest: https://us.sagepub.com/en-us/sam/the-sage-handbook-of-case-based-methods/book230470#contents

Best

Ingo