BSSn4495:
Qualitative
research in
security
studies
The comparative
method
March 8, 2022
Miriam Matejova, PhD
Agenda
• Theory vs methods
• Qualitative vs. quantitative
methods
• The comparative method
Theory vs.
methods
Quantitative vs. qualitative methods
• Ragin: case-oriented vs. variable-oriented
approaches
– Case-oriented researchers: cases as meaningful
but complex configurations of events and
structures; singular, whole entities purposefully
selected
– Variable-oriented research: “homogeneous
observations drawn at random from a pool of
equally plausible selections” (Ragin 2004)
Quantitative vs. qualitative methods (cont.)
• Quantitative methods focus often on theory
testing
– BUT we also need concept creation, elaboration,
refinement
• The issue of conjunctural causation
– in-depth investigations of individual cases can
identify complex patterns of conjunctural causation
Comparative research
• Goals:
– Causal analysis;
– “Parallel demonstration of theory” (i.e., show that
a theory explains the case);
– “contrast of contexts” (i.e., show how different the
cases are; how parallel processes play out in
different contexts)
The comparative method
• When should we use the comparative method?
• Purpose: primarily to test hypotheses; discover
empirical relationships among variables
– Could be used to build new theories
• Good for: addressing spurious correlation
The comparative method (cont.)
• One of the four fundamental methods that
can be used to test the validity of general
empirical propositions (Lijphart 1971)
• Methodology of comparison; a method or
approach, not a technique
• Focus on cases instead of variables alone
• Usually involves small-N research
Most Similar Systems (MSS) design/ Mill’s
method of difference
• Comparing similar cases that show different
outcomes will make it easier to control for
factors that are not the causal agent and
isolate the independent variable that explains
the presence or absence of the dependent
variable.
Most Different Systems (MSD) Design/
Mill’s method of agreement
• Comparing very different cases that all have
the same dependent variable will allow
identification of a point of similarity between
otherwise different cases → identification of
the independent variable that is causing the
outcome.