TAK
I think the point that I and most folks here are trying to make, is that REAL (as in reliable and complete) data is scarce on the ground. Yes there are numbers you can use BUT if these don't reflect reality, all the statistical brilliance in the world won't help. You're basically just shrugging your shoulders with math. Just because you can do a great regression analysis and show that the answer is 47, doesn't make it so. What I was trying to say obliquely, and now will say explicitly, is that maybe a less quantatative approach might be warranted. Be driven by the question not the method. What is your thesis? Tell me that and I can point out the data you need to test it.
FQ13