Dear All,
I am looking for a data set in the form of a contingency table. Ideally I
would like at least a four way table that exhibits a reasonable amount of
model uncertainty. I would prefer large dimensional interactions to be
present in any chosen log-linear model. I currently have the following data
sets.
1. Alcohol, Obesity and Hypertension (Knuiman and Speed 1988)
2. German Political attitudes (Wermouth and Cox 1998)
3. Risk factors for coronary heart disease (Edwards and Havranek 1985)
Set 1 is a 4*3*2 table, thus the largest interaction is 6 dimensional, I
would like larger. There is little model uncertainty in data set 2. Data
set 3 is a 2^6 and thus the largest interaction possible is of dimension 1!
Of course I could generate my own data set, but I would rather have a set
that has appeared in statistical literature.
Could anyone point me towards some good references?
Best Wishes
Roger Gill
|