Dear Students,
1. Just a reminder, the way the CLE is set up, the material you should be working on NOW (the lecture and HW on Chapter 2, dichotomous tests) is posted with next Tuesday's Interactive Large Group session. That session will be much more valuable if you view the lecture and try the problem set first. I hope to see you then!
2. In Vidya's section today someone asked about weighted Kappa for nominal (not ordered) variables. I couldn't think of an example of a nominal variable for which some values might be closer together than others, but Chris Carvalho suggested people tasting and identifying sodas. People should get more credit if one calls a beverage Coke and the other Pepsi than if one says Coke and the other Mountain Dew. (Wine also works -- more part credit if both (blinded) tasters think the wine is red.) Can anyone think of a medical or public health example?
Tom