Dissertation drive

This year I will complete my PhD. Right now I'm working on my dissertation prospectus, or proposal, which explains to my dissertation committee what precisely I intend to do. It will have the following parts:

The literature review - this discusses what other researchers have learned or theorized about my topic

Topic explanation and justification - this will describe what topic I plan to research, and what questions I have about the topic that I will try to answer with my research. It also explains why the research is useful/necessary, and makes it clear that my same questions haven't been asked and answered by anyone else already.

The research plan - this details exactly how I will go about doing the research. It will explain what data I will use, how I will collect the data, how I will make sure the data is "clean" - not biased, how I will "clean" the data if it does have the potential for bias, and what statistical formulas I will use on the data to find out the answers to my questions.

Once I've written this proposal, I'll send it to my dissertation chair and the other committee members. They'll look at it, make comments and suggestions for change, then send it back to me for revisions. I'll revise according to their suggestions, and send it to them again. When they think it's ready, my dissertation chair (hereafter, "Chair") will set up a date for me to make a public presentation on the proposal. That will be at my school, and open to anyone who wants to attend. By the time I get to that point, the committee should have reviewed it often enough to be able to approve the proposal without substantial changes. There's always a possibility another professor or student will make a suggestion that could cause my committee to request big changes, but as far as I know, that rarely happens.

Once the committee has accepted my proposal, the Real Work begins. I have three datasets - Birmingham homicide data for 1991-1992, newspaper articles from the Birmingham News for the same time period, and the 1990 US Census data for the Census grid areas containing the address of the homicide event (not the victim or offender address, but where the murder happened). I will choose the appropriate information from each set, run the statistical analyses I've identified in my proposal, and then analyze the results to see if what I expect to see is what is actually happening.

"What I expect to see" is my theory. The "null hypothesis" is the default to my question. In this case, research indicates that typically news stories are selected based on the rarity of their occurrence. My theory is that journalists select which homicides to cover not based on their rarity, but rather on which homicide stories will most reflect what society already tends to believe about the group(s) involved in the homicide. So the null hypothesis is "journalists select which homicides to cover based on their rarity". The results will either support my theory or will support the null hypothesis. Naturally it is more interesting and fun if my theory is supported, but it is just as important to find that my theory is not supported. The study I am replicating concluded that the null hypothesis was NOT supported and that journalists DO select homicides to cover based on what will reinforce what society wants/tends to believe already.

I'm going to use this blog to report my progress every day that I work on my dissertation. A lot of friends and family have shown interest, and my dissertation group as well. I'll also use it to think out loud about what I'm doing. I find that explaining things to other people often helps me think more clearly about it myself.

Or that's the theory, anyway :).