I stayed at work later than I'd planned, mostly because Metro isn't my personal taxi. I spent a long time trying to figure out whether the right way to determine whether I was supposed to use a likelihood ratio or a Wald test to determine whether estimates from stratified analyses were statistically different from each other. The problem was that the method that I thought was right (lrtest) wasn't giving me the answer that I thought was right. I ended up sending an email to Thomas, though it's probably somewhere in my biostat notes from this fall.
By the time I got home I was starving, which really improved the quality of pre-prepared Indian food.
As much as I'd love to find out whether The Speedles are really English, the FUNC is not easily accessible to carless me at this hour. Instead of hearing punk versions of Beatles songs, I think I'll watch Blue Velvet.