Paradigms and Sand Castles: Theory Building and Research Design in Comparative Politics
Geddes’s piece of work spoke to my heart, I do feel the engine of doing research is out of my curiosity rather than data. But as a freshman in academia, I disagree on how the research process starts. Realistically speaking, without proper data, it’s very difficult to prove my perception of how things work. There is a huge gap between concepts and operational index. For instance, how to indicate a country had transformed from a totalitarian country to a democratic country. There are different types of the index of regime type: freedom house, Polity IV, and etc.. Our current interest is not included in existing database by focusing more on democratic accountability. However, this index exists for a long time, and is established and maintained by prestige research centers. Even if we create our own database, the main obstacle is to convince other why this is much better after sorting out the fund of setting up a database. Therefore, with existing database, the nuanced difference in our theory is not quite obvious.
Though "(multiple regression)is not always a good image to have in mind when trying to explain something complicated" sound plausible in content, it does not provide a better way to build the link between causal factors. I believe people usually have an initial theory in mind, but we tend to add on more and more variables ad hoc to explain the uncovered places. This is the hardest task in comparative studies, cause some time we are not sure about the deviation of cases is insignificant to discard, the criteria just too vague and usually case by case. And it's actually quite useful to use multiple regression to think of the situation before you have a clearer image of how the causal factors work. Truly, this would easily lead into a dark corner when we doubting if all the necessary information had to be gathered to avoid a misreading of the phenomena.
Also, the problem of going with my curiosity and intuition could lead to a risky situation: interpret the data into what I thought is true or over-implication. Sometimes, data from the empirical study or from the field research don’t provide that kind of information, but it fits the dots of the story map in our head, so we link the dots automatically.
In the end, I think the biggest problem is how knowledge accumulation exists. We usually chose the literature which fit our theory or provides a good platform to lay out our theory. It’s usually very subjective, and presenting every piece of previous related literature is just useless. First, it requires lots of hard work. Second, it detours readers’ attention and easily makes them miss the point. So is knowledge accumulation really exist? Sometimes, while two theories are plausible, but in deep heart, you prefer theory A over theory B without credible or available evidence. It’s very easy for authors just cite the theory B in the footnote and provide a “not bother to explain” content and the link of the scholars in theory B. Sometimes, I felt it’s very convenient to do that, but I feel guilty about that. Therefore, in practices, it’s very hard to have the resources to just follow my inquiry without considering other factors.