When it comes to case collection, Gary King, Robert Keohane, and Sidney Verba warn versus "choosing within the dependent variable". They argue for example that researchers can't make valid causal inferences about war outbreaks by only considering instances wherever war did come about (the researcher also needs to look at https://tysondnpwo.tinyblogging.com/not-known-factual-statements-about-hbs-case-study-writing-help-73655836