Consider the SAS program we ran to simulate the following data

y  = specific gravity

x1 = number of fibers per mm^2 of spring wood
x2 = percent springwood
x3 = percent light absorption of summerwood

according to the model

y  = 0.5 - 0.005x1 + 0.005x3 + E

where E denote iid normal random variables with mean=0, sd=sigma=0.02.

============
= EXERCISE =
============
Modify the code so that only one call of the macro "fibergen" is needed
to simulate data from as many experiments as you like under the model.
You'll need to add
1) a do-loop inside the datastep
2) a new macrovariable, say &numexpts, indicating how many times you'd
like to simulate data from the experiment
3) a by-step within each run of PROC REG