Given: Two sets of box plots. Suppose one represents Time 0. The second one represents Time 1. (As a for-instance, perhaps the data are from population surveys, something like Likert scales, prior to a large marketing campaign, and given again following the marketing campaign.) Through SAS Boxplot we are able to create a "history" dataset, which holds the key information essential to the box plots. Thus, here is the partial results of "history" Data Set 0: Combo v50501L v505011 v50501X v50501M v505013 v50501H v50501S v50501N 20201 3.07941 3.16402 3.43257 3.28578 3.56120 4.57473 0.41611 16 20202 3.08643 3.21101 3.85322 3.88925 4.47217 4.66491 0.56595 29 20203 3.09275 3.93011 5.04295 4.88278 6.10475 8.10604 1.35891 51 20204 3.09588 3.59688 4.41027 4.26929 5.07548 6.18251 0.89046 50 20205 3.07901 3.63219 4.43230 4.26111 5.11600 6.70591 1.00551 46 And here is the partial results of "history" Data Set 1: Combo v50508L v505081 v50508X v50508M v505083 v50508H v50508S v50508N 20201 9.7643 10.2730 10.6535 10.6346 11.1293 11.4314 0.5046217883 34 20202 9.8003 10.7847 11.3163 11.5175 11.8450 12.4824 0.7071127263 52 20203 9.7950 11.0608 12.3718 12.4882 13.4592 15.1603 1.4240605479 59 20204 10.1470 11.3915 12.8449 13.0140 14.4063 15.5220 1.7020733552 57 20205 9.7520 9.9268 10.1919 10.1796 10.3164 10.8573 0.3095559921 25 What I'd like to do is get the difference between Data Set 0 and Data Set 1. Plain and simple. Cell for cell. (Never mind that some cells may be meaningless, i.e., the last column, difference in N.) For instance.... Combo 20201, v50508L = 9.7643 and Combo 20201, v50501L = 3.07941 Difference = 6.68489 And on, and on. Can anyone think of a way to do this directly in SAS? If not in SAS, then in any other program? Thanks much! Nicholas Kormanik .
... View more
