Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Re: Re: Re: creating a "years since event" variable in panel data


From   "Sergiy Radyakin" <Radyakin@aoek.uni-hannover.de>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Re: Re: Re: creating a "years since event" variable in panel data
Date   Thu, 8 Mar 2007 21:42:25 +0100

Hi, Christer,

My code does not make use of your "time" variable. I don't know what "time" means. Is this month?

If "time" really means "wave" (time survey is made, typically numbered from 1 though), i.e. data collected in 1994 are assigned time=0, in 1995 time=1, etc untill 2002 time=7 then you have nothing to change in the code.
You can even drop this variable (it is redundant), since you can always recreate it later with
gen time=year-1994

Note that you will have missings (d=.) until the event occurs for the first time.

Best regards,
Sergiy Radyakin



----- Original Message ----- From: "Christer Thrane" <lineea@broadpark.no>
To: <statalist@hsphsun2.harvard.edu>
Sent: Thursday, March 08, 2007 9:12 PM
Subject: st: Re: Re: creating a "years since event" variable in panel data



Thanks! I think I understand it. Does this also hold for the following data structure?
id year time event
1 1994 0 0
1 1995 1 0
1 1996 2 0
..
1 2002 7 0
----
2 1994 0 0
2 1995 1 0
2 1996 2 1
2 1997 3 1
..
2 2002 7 1
----
3 1994 0 1
3 1995 1 1
3 1996 2 1
..
3 2002 7 1
...



----- Original Message ----- From: "Sergiy Radyakin" <Radyakin@aoek.uni-hannover.de>
To: <statalist@hsphsun2.harvard.edu>
Sent: Thursday, March 08, 2007 8:51 PM
Subject: st: Re: creating a "years since event" variable in panel data



Hello Christer,

I assume you have data in the following form:
. l ,sepby(id)

+-------------------+
id year event
-------------------
1. 1 1997 0
2. 1 1998 0
3. 1 1999 1
4. 1 2000 0
5. 1 2001 0
6. 1 2002 1
7. 1 2003 0
8. 1 2004 0
-------------------
9. 2 1997 1
10. 2 1998 1
11. 2 1999 1
12. 2 2000 0
13. 2 2001 0
14. 2 2002 0
15. 2 2003 0
16. 2 2004 0
-------------------
17. 3 1997 0
18. 3 1998 0
19. 3 1999 0
20. 3 2000 0
21. 3 2001 0
22. 3 2002 1
+-------------------+

(Namely, I don't understand what you mean by time ranging from 0 to 7)

Type:

gen d=.
bys id (year): replace d=cond(event==1,0,d[_n-1]+1)

l ,sepby(id)

Which results in the following:
+-----------------------+
id year event d
-----------------------
1. 1 1997 0 .
2. 1 1998 0 .
3. 1 1999 1 0
4. 1 2000 0 1
5. 1 2001 0 2
6. 1 2002 1 0
7. 1 2003 0 1
8. 1 2004 0 2
-----------------------
9. 2 1997 1 0
10. 2 1998 1 0
11. 2 1999 1 0
12. 2 2000 0 1
13. 2 2001 0 2
14. 2 2002 0 3
15. 2 2003 0 4
16. 2 2004 0 5
-----------------------
17. 3 1997 0 .
18. 3 1998 0 .
19. 3 1999 0 .
20. 3 2000 0 .
21. 3 2001 0 .
22. 3 2002 1 0
+-----------------------+

Notice that in this version it is "duration since the most recent occurance of the event". Alternatively it can be "duration since the first occurance of the event".
First is reasonable if your event is "got a new job" -- then d is tenure. Alternatively, if your event is something like "a child is born", -- then dd might be "duration of life with children" (though no account is given for possible move-outs). Then one will need to change the program:
gen dd=.

bys id (year): replace dd=cond(dd[_n-1]==.,cond(event==1,0,.),dd[_n-1]+1)

+----------------------------+
id year event d dd
----------------------------
1. 1 1997 0 . .
2. 1 1998 0 . .
3. 1 1999 1 0 0
4. 1 2000 0 1 1
5. 1 2001 0 2 2
6. 1 2002 1 0 3
7. 1 2003 0 1 4
8. 1 2004 0 2 5
----------------------------
9. 2 1997 1 0 0
10. 2 1998 1 0 1
11. 2 1999 1 0 2
12. 2 2000 0 1 3
13. 2 2001 0 2 4
14. 2 2002 0 3 5
15. 2 2003 0 4 6
16. 2 2004 0 5 7
----------------------------
17. 3 1997 0 . .
18. 3 1998 0 . .
19. 3 1999 0 . .
20. 3 2000 0 . .
21. 3 2001 0 . .
22. 3 2002 1 0 0
+----------------------------+

Best regards,
Sergiy Radyakin





----- Original Message ----- From: "Christer Thrane" <lineea@broadpark.no>
To: <statalist@hsphsun2.harvard.edu>
Sent: Thursday, March 08, 2007 7:23 PM
Subject: st: creating a "years since event" variable in panel data



Hi,

I have a annual panel data set (1 obs per year), where ID identifies the person and the variable OBSYEAR the year of observation (1995=1, 1996=2, ..., 2002=7).

The variable TIME is the time dimension, ranging from 0 to 7.

The variable EVENT is a dymmy (1=event happened during the year in question; 0=event did not happen during the year in question).

Q: How do I create the variable NUMBER OF YEARS SINCE (A POSSIBLE) CHANGE FROM 0 to 1 on EVENT?

Best regards,

Christer Thrane
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index