4. Run the short program below and then run a one-way t test to determine if the difference scores (Diff) come from a population whose mean is zero. In this program, the RAND function is generating uniform random numbers (numbers between 0 and 1 (all with equal likelihood). The DO loop generates 20 of these random numbers and outputs them to the Difference data set. If you don't like to type, this program and all the other programs associated with the exercises are included in the program Create Datasets.sas. *Data set Difference; data Difference; call streaminit (13579); do Subj1 to 20; Diff .6 rand('uniform'); output; end; run; 5. Rerun exercise 4, but change the value of 20 in the DO loop to 200. What do the tests for normality show? Is it OK to use a t test anyway? How did the p-value change?

Fig: 1