*this program calculates regular and bonferroni t critical values.  To use the program enter
alpha--the per comparison error rate if you are controlling the per comparison error rate
     --the upper limit on the family wise error rate if you are controlling the familywise error rate
ntails--1 for a one-tailed test or 2 for a two-tailed test
c--1 if you are controlling the per comparison error rate or the number of tests if you are controlling 
   the familywise error rate. 
n--the sample size
k--the number of betas (plus gammas and deltas, where relevant) in the model from which the results were computed.
In the example
alpha=.05, ntails=2, c=1, indicating a two tail test and control of the per comparison error rate, 
k=4 indicating 4 regression coeffcients in the model;
data;
input alpha ntails c n k;
cval=tinv(1-alpha/(c*ntails),n-k-1);
datalines;
.05 2 1 50 4
proc print;
run;
quit;