As data are coded, theoretical propositions are generated, categories are developed, patterns are detected, and working hypotheses are established. Through theoretical sampling, more cases are added and more observations and connections are made, which are then compared to the previously developed working hypotheses. Initially, new cases are added based on their similarity to earlier cases. When no new insights are forthcoming from the similar cases, the researcher harnesses dissimilar types of cases. This iterative process is repeated until no new insights are forthcoming. The overall cycle continues until the researcher is satisfied that further recruitment will not change the findings.