In addition to doing my taxes (ugh), I also need to pack this afternoon for the New Horizons in International Conference at the University of Illinois. I'm leaving tomorrow with three fellow U-Iowa students and will present my paper "Dissent in Digital" on Saturday.
I have one major concern with the paper in its current state--reliance on "dummy" variables. A dummy variable is basically a binary variable: 1 equals what you're measuring, 0 equals everything else. For example, one of my variables measures levels of minority religious discrimination. To make the data fit into the mathematical model, rather than measure the scale of discrimination I had to convert the variable to 1 = minority is discriminated against and 0 = minority is not discriminated against. Thus I lose variable fidelity.
That said, I'm hoping for good feedback for improving my model; and even given my concern, I still think I'm on to something...
If you're interested, here's a link to my latest draft. Please feel free to leave comments here, or write me at james -at- jdfielder- dot -com
-
Comments