Friday, 29 May 2009

Unit of analysis

What's your unit of analysis?

That was the coffee time question posed by our Resident Questioning Student. He often starts interesting discussions. I'd naïvely thought it a simple answer, but the more he posed questions, the more complicated the answer became.

The unit of analysis fits in with and stems from your research question. It governs the way you then design your research because it governs what kind of information you seek. That information builds up to give a central picture. I have a systems understanding of the unit of analysis; to me a unit of analysis must have a central focus, a purpose that all its components contribute to, and a unit of analysis consists of a number of components, any one of which, if it were removed, would create a different system. For example, RQS's unit of analysis concerns a process (a campaign) that takes input and outputs x or y, and along the way, there are regulatory processes, decision making processes and ethical processes that influence the outcome. If you took away the regulatory process, you'd have a different system (like the UK MPs could have a different expenses system if their regulatory process was different).

The B852 course that we did for the MRes gave us some useful readings on unit of analysis, especially the Block I papers by Michael Crotty on the research process, and the Bryman & Bell paper on research designs.

RQS is writing his methodology chapter, and warns us that it is pivotal to all your research, which is probably why my third party monitor suggested I start writing it now, this year.

No comments: