USING SAS ON OLYMPUS1/UNIX WORKSTATIONS IN CC2   sta5205  Hoffman 1/24/01

- To Login
  o wiggle the mouse till a window appears
  o it will be 1 of 2 windows.  
   If   "Welcome to..." then type in your username and password
   If   "Select a server..." then click on options and return-to-logon
  o On boot up it may ask which choice of Desktop you want; select CDE
  o Windows will now open; including a director tool bar at the bottom and a file manager.  
You may want to use the file manager to create a new file/directory called sas at this time
  o To logout, click on the right button in a blank area of the screen and select logout
- Usage Hints
  o click on a windows square to enlarge; click on a windows dot inside square to hide; 
click on the arrow for options
  o click on the director toolbar at the bottom to see available programs
- Get to the Web
  o click on the tools drawer in the director toolbar at the bottom (3rd icon from right)
  o click on Desktop Apps and scroll down
  o click on web browser
  o enter our class website in the blank box by typing
http://pegasus.cc.ucf.edu/~hoffman/sta5205.html
view the materials there
- Using Lenth's sample size calculator
  o click on the calculator's website at our class website and if java works correctly 
the site will load in about 1 minute
  o click on ANOVA and click the run button
  o click on the built-in models to select one 
  o change the numbers in the levels box to correspond to the experiment you are doing
  o enter the number of observations you suspect you might use; you can try different values 
in an iterative scheme to ascertain sample size for required power
  o click the differences/contrasts button (only press once and then wait a moment)
  o as a general rule, select the Scheffe method and stick with the -1 1 contrast and by c
clicking on the little box in the top right corner of the detectable contrast slidebar you 
can enter your desired contrast difference of  delta 
  o enter your desired alpha level
  o enter your estimate of the SD(within) (i.e. the standard deviation)
  o read the power off the bottom right slidebar.
can you get r=21 like we did in class for the bead experiment when delta   is 50 the stdev
   estimate is 44 for  alpha
 = .05 and power of .90?  when you finish this return to the sas code 
website of Dean and Voss and leave your web browser open
- Opening SAS
  o click on the tools drawer
  o click on the Desktop
  o click on Terminal
  o enter SAS at the > prompt and wait
- Using SAS
  o you can type code into the editor window or copy it in.
  o to copy code first go to the ran.sas page in your web browser and on the top toolbar click 
Edit-Select All-Copy
  o return to your SAS editor by clicking in that window to activate it and then c
click on the top toolbar on Edit-Paste text
  o to alter our copied program note that a 'd' on the line number 
and <CR> (hit enter key) deletes a line; an 'i' <CR> inserts one.  you may type on new lines as usual
Can you change the number of random units to be generated?
  o end each program with:  run; quit;
  o to execute the SAS code, click on Local-Submit
  o pull up your Log window to see if there were any errors during the run
  o your output appears in the Output window...a compact view can be gotten via clicking 
on Edit-Options-Page
  o to see your code again, click on Local-Recall text in the editor window
  o to save your file click on File-Save As and if you created a sas directory, 
type sas and <CR> to get there, now give your file a name and click OK
  o to Open a saved file, click on File-Open
run your ran.sas a few times
- Illustration of the running of ANOVA and follow-ups via the bead example
  o in the SAS editor copy in battery2.sas from the Dean and Voss website
  o delete the battery data and put in our data:
1 100
1 93
1 140
2 105
2 45
2 85
3 35
3 40
3 37
  o change battery words to size and time words
  o ESTIMATE gives information for confidence intervals and CONTRAST does tests.
  o MEANS statement alone gives estimates of groups sample means along with standard
 deviations. MEANS with your favorite multiple comparison method will 
allow you to do the classic follow-ups. Note that the code below is for illustrative 
purposes only since we would not do all pairwise AND 
control vs. treatment etc.(that would elevate our overall experimentwise alpha level). 
Below mimics what we did by hand in class.
  means size;
  estimate 'svl'
          size   1  0  -1  ;
  estimate 'svm+l'
          size 2 -1 -1/divisor=2;  /* talked about in class, i.e. small versus average of med and large */
  contrast 'svl1'
          size      1 0 -1;
  means size    / bon   cldiff alpha=.15;  /* this is .15 to get .15/6 = .025 -- what we used in class */
  means size    / tukey cldiff alpha=0.10;
  means size    / dunnett('3') cldiff alpha=.10; /* the 3 tells which is the treatment */
  means size    / dunnettl('3') cldiff alpha=.10; /* gives one-sided intervals */
  means size    / scheffe cldiff alpha=.10;