Reflection of theme 4

What I have learnt this week is that there certainly are many problems, both with qualitative methods as well as quantitative ones. Since the paper I picked last Friday used automated information gathering, I didn't delve into the problem aspect of questionnaire bias in a method that I considered quantitative (although I did raise other issues in this paper). My vision of quantitative methods being devoid of (what I had previously considered exclusively) qualitative problems regarding feelings and opinions was shattered.

For instance, by looking at the example of questionnaire answers I realize one must consider qualitative aspects of how people reply to these things. Although I have both answered and created a few questionnaires, I haven't really thought much about how different personalities or individuals get overrepresented certain studies since they are more willing to take the time to answer a questionnaire. And so bias by not participating or by overrepresentation of certain groups of people is certainly an issue. But the same can of course be said about qualitative methods in how some persons are more willing to e.g. attend an interview.

This week we discussed a bit about how theory comes first and quantitative study last to 'prove' your theory. For instance, one might have an idea of how something worked, but wants confirmation by sending out a questionnaire to answer. But how does one come up with this questionnaire so it doesn't limit the answers or affect the outcome in its sheer design? In Wednesday's lecture it was mentioned how questionnaires definitely should be tested before they are sent out. So how do you test a survey before you send it out? Send it out twice, or send it out to another huge group of people? The answer is by discussing with other people, interviewing them, etc. in a smaller more manageable context - i.e. in a qualitative way.

In my paper, the 'theory' consisted of a hypothesis or thought that 'microblogging' is not an entirely new idea in today's society, and in a way proved it by classifying thousands of tweets in historical categories. But it also raised new questions and brought to light new categories or classifications that the authors didn't think of. Because of this, the paper itself was limited in how the predetermined data categories formed the result, and didn't pick up on data that was not in this state space.

This is why one could view qualitative methods in another light: to gain new insights, develop theory, etc. Especially if theory is more or less just hypothesis based on research on 18-19th century diaries (like in my paper) - it could very well be developed by interviewing and discussing with people using Twitter today, for instance. By looking at the two methods in this way, one does not have to discuss quantitative vs. qualitative methods as if the one excluded the other, but rather in how they can best be utilized together.