"Real life in real time is lots of noise and not very much signal."

Luker discusses sampling - the important practice of collecting data. One of the very risky problems with sampling is choosing such a segment of data that clearly comes out as influential to the research outcome.  I thought of how often we come upon studies or research conclusions in everyday media that are so obviously biased.

 

On the INF 1240 blog, Professor Grimes posted an analysis of the study about a correlation between divorce rates and shared housework. Describing the problems with the study, Professor mentions various degrees of bias, with which I very much agree. In her blog, Professor writes that it is important to always ask questions about research and researchers. This example with divorcing couples who share housework is particularly good to describe poor and incomplete ways of collecting and then presenting data. This unsupported study clearly creates a sensationally politicized outcome, which may be its only purpose, really: how many people will have an emotional response to such a finding about sharing housework! You only hope that with pieces of information like this, people do stop and ask questions.

I find that biases at times are obvious, but also at times obscure. Luker talks about sampling at length, which, perhaps, shows that carefully, inoffensively, and properly accessing data for research is not all that easy. It is clear that many researchers struggle with choosing the right place and the right ways to locate data. I find the three guidelines she uses very helpful: find a setting where variable varies, make sure the part stands for the whole, and let theory lead your sampling. I ended up questioning my own methodology in my research proposal, after reading this. In my research I initially wanted to interview my own peers who are in the field I am studying. This, of course, goes against the two of the three guidelines. There are truly many things to be careful about.