Prev: proc gmap, strange rotation at random
Next: Error Handling – Data Sets
From: Ted Ames on 1 Dec 2009 16:11
Final Reading

Normal Mild Moderate Severe Missing Total
Normal
Mild
Moderate
Severe
Missing
Total

Sorry, my previous Table Shell got "jumbled". The Final Readings are on top
and the Baseline readings are to the left of the column.
From: Arthur Tabachneck on 1 Dec 2009 16:34
Ted,

Sure sounds a lot like the table produced/described in the following paper:
http://www.lexjansen.com/pharmasug/2003/posters/p111.pdf

HTH,
Art
-----------
On Tue, 1 Dec 2009 16:11:03 -0500, Ted Ames <forwardwash(a)YAHOO.COM> wrote:

>Final Reading
>
>Normal Mild Moderate Severe Missing Total
>Normal
>Mild
>Moderate
>Severe
>Missing
>Total
>
>Sorry, my previous Table Shell got "jumbled". The Final Readings are on
top
>and the Baseline readings are to the left of the column.
From: Ted Ames on 1 Dec 2009 19:02
Forgot to mention that each "cell" is the number of customers for each
baseline-final value combination. There are 2 totals. A row total as well as
a column total. Sample data might look like this:

customer Baseline Final
1 Normal Mild
2 Severe Normal
3 Normal Normal
4 Mild Severe


decodes:
Normal=1
Mild=2
Moderate=3
Severe=4
Missing=99
From: Ted Ames on 1 Dec 2009 20:28
I have to "zero fill" the cells so even if a combination does not occur I
have to 0 fill it.
From: NordlDJ on 1 Dec 2009 20:44
> -----Original Message-----
> From: SAS(r) Discussion [mailto:SAS-L(a)LISTSERV.UGA.EDU] On Behalf Of Ted
> Ames
> Sent: Tuesday, December 01, 2009 4:02 PM
> To: SAS-L(a)LISTSERV.UGA.EDU
> Subject: Re: Creating a "Shift" Table
>
> Forgot to mention that each "cell" is the number of customers for each
> baseline-final value combination. There are 2 totals. A row total as well as
> a column total. Sample data might look like this:
>
> customer Baseline Final
> 1 Normal Mild
> 2 Severe Normal
> 3 Normal Normal
> 4 Mild Severe
>
>
> decodes:
> Normal=1
> Mild=2
> Moderate=3
> Severe=4
> Missing=99

I am not sure what form you want your final output to be, but this is basically just a PROC FREQ using the sparse option. Each value must be in the data for baseline and final at least once (but not all combinations are necessary) for the sparse option to do its work. Something like this should get you started:

data have;
input customer Baseline $ Final $;
cards;
1 Normal Mild
2 Severe Normal
3 Normal Normal
4 Mild Severe
5 Moderate Moderate
;
run;
proc freq data=have;
table baseline*final / sparse nopercent nocol norow ;
run;

If you need something else, then write back to SAS-L with more detail on your wanted output.

Hope this is helpful,

Dan

Daniel J. Nordlund
Washington State Department of Social and Health Services
Planning, Performance, and Accountability
Research and Data Analysis Division
Olympia, WA 98504-5204
 |  Next  |  Last
Pages: 1 2
Prev: proc gmap, strange rotation at random
Next: Error Handling – Data Sets