What an elegant solution. After checking back over several and finding no repsonses (I'm new to the group and have not figured out how to see complete thread), I worked 2 days, wrote over 300 lines of code and was finally able to generate the table. You demonstrated how simple this was. Thank you.

data _null_, wrote: >I like this a little better. I believe it is more similar to the OPs >original need. > >title 'Balanced Data from Randomized Complete Block'; >data plants; > input Type $ @; > do Block = 1 to 3; > input StemLength @; > Stem2 = StemLength+rannor(12345)*4.1; > Stem3 = StemLength+rannor(12335)*.4; > output; > end; > datalines; >Clarion 32.7 32.3 31.5 >Clinton 32.1 29.7 29.1 >Knox 35.7 35.9 33.1 >O'Neill 36.0 34.2 31.2 >Compost 31.8 28.0 29.2 >Wabash 38.2 37.8 31.9 >Webster 32.5 31.1 29.7 >;;;; > run; > >ods output mclines=mclines; >proc glm order=data; > class Block Type; > model Stem: = Block Type / solution; > means Type / waller regwq; > run; > quit; >data work.mclines2; > set work.mclines(where=(mean ne ._)); > array line[*] line:; > call sortC(of line[*]); > length line0 $8; > line0 = lowcase(cats(of line[*])); > if _n_ eq 1 then call symputX('F',cats('$F',dim(line),'.')); > run; >proc report nowd list headline data=work.mclines2; > column level method,dependent,(mean line0) necessaryEvil; > define level / group; > define method / across 'Method' '--'; > define dependent / across 'Dependent' '--'; > define mean / sum ' ' format=6.3; > define line0 / display ' ' format=&f; > define necessaryEvil / noprint; > compute after; > line 'Means with the same letter are not significantly different.'; > endcomp; > run; > >> I found data in the examples that should be pretty close to yours. >> This may be something like you need. >[quoted text clipped - 58 lines] >> > I can get everything except the significance (charcter data). Is this even >> > possible?

-- Message posted via http://www.mathkb.com