I find it interesting how Miles stated that case studies are not rational and therefore they are less scientific. Despite this view Yin points out that case studies can be conducted by either qualitative or quantitative methods. Yet, mixed-methods have been used in case studies, which yields positive results. If case studies are thought to be so unscientific then why are researchers using them? What could they possibly offer the field that the so-call scientific methods cannot? This reminds me of something Luker said about how canonical social scientists do not like methods that are unscientific in nature, which is why they do not study the phenomenon but rather the population being sampled.
Case studies can be explanatory, descriptive, and exploratory. It is safe to assume that based on this idea of what case studies are is why some researchers consider them to be unscientific. When one conducts several case studies he or she is looking for patterns amongst them to see if they have anything in common, which enables a well founded explanation for why or why not things seem to be this way. Case studies can be conducted using participation observation or ethnology but it can also use other methods to collect and analyze the data. They are beneficial for collecting data that requires open-ended questions that is limited to a particular research intent.
I like how Yin uses a detective analogy as a comparison to how a researcher would conduct a case study. Based on the data collected the researcher may have to adjust the theory or framework of the case study. The research tailors his or her decisions based on the relevancy of the data. In order to do so the researcher needs to take good notes and after the interview or observation is done he or she should quickly write down everything that took place to the best of his or her memory or else he or she will forget later down the road. This helps when one is completing the final research product, especially if it is several months after the data was collected.