/* option 1 - using CI calculation in PROC MEANS */
data junk;

do iexpt=1 to 10;
  do i = 1 to 50;
    y = RAND('Normal');  * here mu=0 as default;
	output;
  end;
end;
run;

proc means data=junk;
  var y;
  class iexpt;
  output out=mysummary lclm=ylclm uclm=yuclm;
run;

proc print data=mysummary;
run;

data calc_cover_length;
  set mysummary;
  if iexpt NE .;
  icover = (ylclm <= 0 <= yuclm);
  yL = yuclm - ylclm;
run;

proc print data=calc_cover_length;
run;

/* option 2 - doing all the work in the DATA step */
data junk2;
array x(50) x1-x50;
tmult = quantile('T',.975,49);
mu = 0;
do iexpt=1 to 10;
  do i = 1 to 50;
     x[i] = RAND('Normal');
  end;

  xbar = mean(of x1-x50);
  xvar = var(of x1-x50);
  xlcl = xbar - tmult*sqrt(xvar/50);
  xucl = xbar + tmult*sqrt(xvar/50);

  icover = (xlcl <= mu <= xucl);
  xl = xucl - xlcl;
  output;
end;
run;

proc print data=junk2;
  var xbar xvar xlcl xucl icover xl;
run;