Date: Mon, 21 Oct 2002 09:46:10 -0500
Reply-To: Carol Albright <calbright@VISI.COM>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Carol Albright <calbright@VISI.COM>
Subject: Re: syntax question-- repeating calculations'
Content-Type: text/plain; charset="iso-8859-1"
Hi, All
I'm resending this because some people may not have been able to read it,
especially if they use GroupWise (I had copy & pasted the syntax & output,
which preserves the formatting but the mixed fonts confuses some e-mail
packages).h
<Unformatted message follows>
First you need to compute a variable which identifies the episode for each
client, say "EPISODE" (make sure this does what you wanted -- I'm not a
whiz with the do if/lag stuff):
sort cases by clientid (a) date (a) ep_ct (a).
compute episode = 1.
do if clientid = lag(clientid) and ep_ct = lag(ep_ct).
compute episode = lag(episode) + 1.
else if clientid = lag(clientid) and (ep_ct - lag(ep_ct)) < 0.
compute episode = episode + 1.
else if clientid = lag(clientid) and (ep_ct - lag(ep_ct)) > 0.
compute episode = lag(episode).
else if clientid <> lag(clientid).
compute episode = 1.
end if.
list clientid date ep_ct episode.
List
CLIENTID DATE EP_CT EPISODE
161.00 01/01/01 1.00 1.00
161.00 02/01/01 2.00 1.00
161.00 03/01/01 3.00 1.00
161.00 04/25/01 1.00 2.00
161.00 05/25/01 2.00 2.00
161.00 06/25/01 3.00 2.00
175.00 01/01/02 1.00 1.00
175.00 02/01/02 1.00 2.00
175.00 03/01/02 1.00 3.00
175.00 04/01/02 1.00 4.00
Number of cases read: 10 Number of cases listed: 10
Then use the handy AGGREGATE command to summarize your data as follows:
AGGREGATE
/OUTFILE='C:\Data/myfilename.sav'
/BREAK=clientid episode
/tx_amts 'Total Treatment Cost per Episode' = SUM(tx_amount).
This will create a new file, myfilename.sav, which includes variables:
ClientId, episode, and tx_amts, which is the total amount for the episode
for the client.
Say you wanted a summary of how much the client spent, regardless of
episode. In that case, change the BREAK to only include clientid.
You can also compute other statistics using AGGREGATE, such as the minimum
or maximum value or the mean. IE:
AGGREGATE
/OUTFILE='C:\Data/myfilename.sav'
/BREAK=episode
/avg_amte 'Mean Treatment Cost per Episode' = MEAN(tx_amount).
would give the mean amount a client spent for each episode of care.
Carol
At 03:04 PM 10/9/02 -0700, Student wrote:
>I am stuck on a syntax problem, and hoped you could help. Clients
>(client ID) received TX (TX Amount) at varying times (some episodes had 1
>month, others more than 1 month). The EP_CT below delineates the episodes,
>starting with the number 1 (i.e., episode 1 for client 161 consisted of
>new# 1, 2, and 3; episode 2 had 4 months (1 to 4); and episode 3 had 3
>months). What I want is a case for each client-episode that totals the
>Tx_amounts from each of the months in that episode. I have attempted
>looping and Lag counting, but have been unsuccessful-- perhaps there is an
>easier way? Any help would be appreciated.
>
>
>TX_ID Mo_Yr ClientID EP_CT Tx_Amount
>103 11/1/1999 161 1 11
>103 12/1/1999 161 2 31
>103 1/1/2000 161 3 3
>1220 2/1/2001 161 1 20
>1220 3/1/2001 161 2 31
>1220 4/1/2001 161 3 30
>1220 5/1/2001 161 4 6
>1220 9/1/2001 161 1 5
>1220 10/1/2001 161 2 1
>1220 11/1/2001 161 3 2
>289 1/1/2001 164 1 0
>289 2/1/2001 164 2 0
>289 6/1/2001 164 1 0
>289 10/1/2001 164 1 0
>
>I want it to look something like this...
>Client_ID Tx_Amnt Episode
>161 45 1
>161 87 2
>161 8 3
>164 0 1
>etc....
>
>
-------------------------------------------------------------------------
Carol L. Albright, MS | E-Mail : calbright@visi.com
Albright Consulting | Phone : 651/699-7218
St. Paul, MN 55105 USA | Research data services
http://www.tc.umn.edu/~syzygy
-------------------------------------------------------------------------
