what is method of agreement
- if a phenomenon occurs in two or more places the explanation must lie in common features
method of difference
- if two or more situations are similar but phenomenon only exists in one of them its cause must be related to a different or two aspect
what is most different systems design (MDSD)
select cases that are very different but share a characteristic we want to study and a common outcome
what is most similar systems design (MSSD)
suggests looking at countries with a lot in common but dont have the same phenomenon or outcome we want to explain
what are high level categories
universal concepts that can be used for comparison worldwide (political systems)
what are medium level categories
general concepts that are used to compare reasonably similar things (left wing parties to cover everything from communist to centrist left)
what are low level categories
for country by country comparisons
what is the second form of selection bias
- arises in qualitative studies that rely on historical sources that may fit the theory being tested
- historic sources are often biased and highlight some events more importantly
third form of selection bias
selecting contemporary time periods and drawing inferences about longer term processes
how can selection bias be avoided
- you know best, are in the news/ are happening currently
- easily prove your hypothesis (not only these though)
- transparent case selection criteria
- include negative cases
- random or stratified sampling for quant
give two examples of case selection bias when studying democratic stability and backsliding
- only study cases where was stability or backsliding, not selecting cases where there was not
- only selecting extreme cases such as Russia or Hungary and not less extreme cases
