5: Rethinking comparisons

Alternative ways off comparing

logic of comparisons that are not motivated by control

1. Focus on similar shared dilemmas faced across seemingly divergent contexts

   • identifying non-obvious political connections across does not equal context

2. Look at linkages & diffusion of practices

Internal Validity

• explanations holds true (is robust) for the cases actually studied (sample)

• measured by how many other things that we need in order to be convinced by that explanation

• how confident are we that the explanation of that case explains that case

External validity

• explanations holds true (is robust) for the cases not studied (population)

• how well does our explanation travel and how far

• there is a trade off between the validities

How can controlled comparisons best generate external validity?

3 guiding principles:

• Select cases that maximise control over alternative explanations

• Compare cases that capture representative variation in population

  • typological representativeness

• findings expressed as general variables/mechanisms instead of context specific