EvilPettingZoo97@AOL.COM replied: >On Sat, 8 Sep 2007 20:09:00 -0000, jason.s.turner@GMAIL.COM wrote: > > >Hi, I am fairly new to SAS and have ran into a little trouble. My > >initial dataset consited of grades (4,3,2,1) and majors. EX: > > > >2 ACCOUNTING > >3 ACCOUNTING > >4 ACCOUNTING > > > >I want to generate a KDE graph of a range of 0 to 4 on the y axis and > >with the majors on the x axis. To do this I did a Proc Means on the > >grade (by major) and got a dataset like this: > > > >ACCOUNTING 2.952293578 > >AG BIOTECH 2.7075471698 > >AG ECON 1.9034482759 > >AG EDUC 2.0147058824 > > > >The problems I am having are #1 when I run Proc KDE, I get no graph > >and #2 all of the charts that come up are per major (when I eventually > >get a graph, I want it all lines for all majors on one graph). > > > >This is my code: > > > >proc sort data = sasuser.impw_0000 out = ukmathsort; > > by major; > >run; > > > >proc means data = ukmathsort mean; > > var grade; > > by major; > > output out = ukmathmeans MEAN = meangrade; > >run; > > > >proc kde data=ukmathmeans; > > univar meangrade/gridl=0 gridu=25 out=kdemath; > >run; > > > > > >Thanks, > > > >Jason > > >Jason, > >If you have grades with just four values, then I don't think you want a >density plot. These are 'smoothed histograms' of continuously measured >parameters. You can do it, but it may not be appropriate. > >To get a graph, invoke the ODS Statistical Graphics that became available >in V9. > >Below is some sample code: > >data scores ; > do major=1 to 50 ; > do _n_=1 to 40 ; > grade=round( ( ranuni( 90210 )*100 )/ 25 ) ; > output ; > end ; > end ; > stop ; > run ; > >ods graphics on ; >ods rtf file="%sysfunc(pathname(work))\junk.rtf" style=journal ; > > proc kde data=scores ; > univar grade / plots=hisdensity ; > run ; > >ods rtf close ; > > >If the stats peeps disagree and think creating densities for discrete >values >is OK, see a previous post that generates overlayed density plots with the >Graphics Template Language (GTL) >http://listserv.uga.edu/cgi-bin/wa?A2=ind0605D&L=sas-l&P=R8878 > >HTH, >Ken

The 'stats peeps', huh? I'll check and see if my stats homeys wanna put a cap in your survivor function. :-) :-)

I agree with you. I think that Jason would benefit greatly from explaining what he is *really* trying to do with these data. At that point, someone here is more likely to be able to point him in a better direction.

HTCT, David -- David L. Cassell mathematical statistician Design Pathways 3115 NW Norwood Pl. Corvallis OR 97330

_________________________________________________________________ Discover sweet stuff waiting for you at the Messenger Cafe. Claim your treat today! http://www.cafemessenger.com/info/info_sweetstuff.html?ocid=TXT_TAGHM_SeptHMtagline2