Kurt, You rock!  Thanks for taking the time to go through this both times.  You have saved me hours of work for each time this survey is collected.  ... View more

Thanks, Kurt. That did something, but I had to drop 'n' from the by statement 'n sch' statement to go anywhere.  I did get through all the data steps, but all the QA1, QA2,...QD1, QD2,.. report out at 100%. If I delete (where=(answer ne 'O')), then I get 27 columns with 100%, so if I could get the correct %s, for A,D,O, then i could use a formula, or drop the O columns. I think something essential is lost in the shift from Long to Long2 (short snip below), but I don't see it. Data Long                   DATa long2                 sch question _LABEL_ COL1 COL2 COL3 COL4 COL5 COL6   sch question _LABEL_ COL2 COL3 COL4 COL5 COL6 answer sch1 Q1 Q1 5 5 3 5 5 5   sch1 Q1 Q1 5 3 5 5 5 A sch1 Q2 Q2 4 5 2 5 5 5   sch1 Q2 Q2 5 2 5 5 5 A sch1 Q3 Q3 5 5 4 5 5 5   sch1 Q3 Q3 5 4 5 5 5 A sch1 Q4 Q4 4 5 2 4 5 4   sch1 Q4 Q4 5 2 4 5 4 A sch1 Q5 Q5 4 5 3 5 5 5   sch1 Q5 Q5 5 3 5 5 5 A sch1 Q6 Q6 4 5 2 4 5 5   sch1 Q6 Q6 5 2 4 5 5 A sch1 Q7 Q7 4 5 2 5 5 5   sch1 Q7 Q7 5 2 5 5 5 A sch1 Q8 Q8 4 5 2 5 4 5   sch1 Q8 Q8 5 2 5 4 5 A sch1 Q9 Q9 4 5 2 5 5 5   sch1 Q9 Q9 5 2 5 5 5 A sch2 Q1 Q1 5 1 5 4 4 3   sch2 Q1 Q1 1 5 4 4 3 A sch2 Q2 Q2 5 1 5 4 4 2   sch2 Q2 Q2 1 5 4 4 2 A sch2 Q3 Q3 5 4 5 4 5 4   sch2 Q3 Q3 4 5 4 5 4 A sch2 Q4 Q4 4 4 4 4 4 2   sch2 Q4 Q4 4 4 4 4 2 A sch2 Q5 Q5 4 5 4 4 4 4   sch2 Q5 Q5 5 4 4 4 4 A sch2 Q6 Q6 4 4 4 4 5 2   sch2 Q6 Q6 4 4 4 5 2 A sch2 Q7 Q7 2 1 5 3 4 2   sch2 Q7 Q7 1 5 3 4 2 D sch2 Q8 Q8 2 1 5 4 4 4   sch2 Q8 Q8 1 5 4 4 4 D sch2 Q9 Q9 4 4 4 4 5 4   sch2 Q9 Q9 4 4 4 5 4 A ... View more

Chris, Thanks for asking.  Yes, I actually want to get 18 columns for each school. 1 for each question with % 1or2 and 1 with %4or5. This code gets me 9 crosslists with the percentages I need, as in the initial post.  When I try to merge all the crosslists, with the code below, I get the output below. data ghost2; set ghost; ans1 = " "; if Q1 ne . and Q1 le 2 then ans1 = "D" ; else if Q1 ne . and Q1 ge 4 then ans1 = "A" ; else if Q1 ne . and Q1 eq 3 then ans1 = "O" ; ans2 = " "; if Q2 ne . and Q2 le 2 then ans2 = "D" ; else if Q2 ne . and Q2 ge 4 then ans2 = "A" ; else if Q2 ne . and Q2 eq 3 then ans2 = "O" ; ans3 = " "; if Q3 ne . and Q3 le 2 then ans3 = "D" ; else if Q3 ne . and Q3 ge 4 then ans3 = "A" ; else if Q3 ne . and Q3 eq 3 then ans3 = "O" ; ans4 = " "; if Q4 ne . and Q4 le 2 then ans4 = "D" ; else if Q4 ne . and Q4 ge 4 then ans4 = "A" ; else if Q4 ne . and Q4 eq 3 then ans4 = "O" ; ans5 = " "; if Q5 ne . and Q5 le 2 then ans5 = "D" ; else if Q5 ne . and Q5 ge 4 then ans5 = "A" ; else if Q5 ne . and Q5 eq 3 then ans5 = "O" ; ans6 = " "; if Q6 ne . and Q6 le 2 then ans6 = "D" ; else if Q6 ne . and Q6 ge 4 then ans6 = "A" ; else if Q6 ne . and Q6 eq 3 then ans6 = "O" ; ans7 = " "; if Q7 ne . and Q7 le 2 then ans7 = "D" ; else if Q7 ne . and Q7 ge 4 then ans7 = "A" ; else if Q7 ne . and Q7 eq 3 then ans7 = "O" ; ans8 = " "; if Q8 ne . and Q8 le 2 then ans8 = "D" ; else if Q8 ne . and Q8 ge 4 then ans8 = "A" ; else if Q8 ne . and Q8 eq 3 then ans8 = "O" ; ans9 = " "; if Q9 ne . and Q9 le 2 then ans9 = "D" ; else if Q9 ne . and Q9 ge 4 then ans9 = "A" ; else if Q9 ne . and Q9 eq 3 then ans9 = "O" ; run; proc freq data=ghost2; tables sch*ans1 /crosslist out=crlist1 nocol nofreq nopercent; run; proc freq data=ghost2; tables sch*ans2 /crosslist out=crlist2 nocol nofreq nopercent; run; proc freq data=ghost2; tables sch*ans3 /crosslist out=crlist3 nocol nofreq nopercent; run; proc freq data=ghost2; tables sch*ans4 /crosslist out=crlist4 nocol nofreq nopercent; run; proc freq data=ghost2; tables sch*ans5 /crosslist out=crlist5 nocol nofreq nopercent; run; proc freq data=ghost2; tables sch*ans6 /crosslist out=crlist6 nocol nofreq nopercent; run; proc freq data=ghost2; tables sch*ans7 /crosslist out=crlist7 nocol nofreq nopercent; run; proc freq data=ghost2; tables sch*ans8 /crosslist out=crlist8 nocol nofreq nopercent; run; proc freq data=ghost2; tables sch*ans9 /crosslist out=crlist9 nocol nofreq nopercent; run; data allcrlist; merge crlist1 crlist2 crlist3 crlist4 crlist5 crlist6 crlist7 crlist8 crlist9 ; by sch; run;   sch ans1 COUNT PERCENT ans2 ans3 ans4 ans5 ans6 ans7 ans8 ans9 sch1 A 5 20.83333 A A A A A A A A sch1 O 1 4.166667 D A D O D D D D sch2 A 6 25 A A A A A A A A sch2 D 2 8.333333 D A D A D D D A sch2 O 1 4.166667 D A D A D O D A sch3 A 2 8.333333 A A A A A A A A sch3 D 2 8.333333 D D D D D D D D sch3 D 1 4.166667 D D O D D D D D sch4 A 4 16.66667 A A A A A A A A sch4 D 1 4.166667 D D D D D D D O sch4 D 1 4.166667 D O D O D D D O sch5 A 3 12.5 A A A A A A A A sch5 A 1 4.166667 D O D A D D D A sch5 A 1 4.166667 D O O A D D O A ... View more

Thanks for asking, but the data is protected and would be difficult to mask so it could be posted.  I was able to use the links to the articles to find statements I could use to get the results I needed. ... View more

Thank you!  That makes it easier.  I was able to run each pair once, and sort the outputby type and group in excel to get a simple chart of each pair by group. The revised code is below. proc corr data = alltoclus outp=OutCorr1 NOPRINT; var var1 var2; by Grade; run; proc corr data = alltoclus outp=OutCorr2 NOPRINT; var var3 var4; by Grade; run; proc corr data = alltoclus outp=OutCorr3 NOPRINT; var var5 var6; by Grade; run; proc corr data = alltoclus outp=OutCorr4 NOPRINT; var var7 var8; by Grade; run; proc corr data = alltoclus outp=OutCorr5 NOPRINT; var var9 var10 ; by Grade; run; ... View more