Pre-generated values of X, drawn from various distributions below once, and then fixed for all uses of 2k1-in-silico. For n observations, use the first n rows; for variable X1, X2 etc. use column 1, 2 and so on.
