In the interest of sanity, and to reduce signal-to-noise distortion, I have decided to focus on one particular industry in my particular state, figure out how to conduct a survey of compliance to a federal law when there is already concurrent state law that is more generous in leave benefits, and divide the survey into organizations of 50-249 employees, and 249+ employees. I will interview the management and when I can have access, the employees. Thus, the industry, human capital requirements, organizational environments, geographical area, and laws are all relatively homogenized factors, so I can focus on organizational size, structure, culture and gender composition to explain differing levels and typology of compliance.
This is a dissertation, people. I don't have the time or money to conduct a nationwide survey of hundreds of different types of organizations in different industries. Let's wait until I get tenure-track and an NSF grant for that one.
For now, I am happy that the parameters are roughly defined and manageable so that I can begin a pilot study in the spring and the real fieldwork in the summer.
So much of a dissertation is having a refined, feasible research question and methodology and not spending all of your time on the literature review (which can be endless if you don't stop at some point). So much of an empirical study is setting limits so that you can actually collect some useful data to work with, so that you can actually work and not sit there staring blankly at the screen.