How about this:
clear
input id wave str7 empstat exit longspl
1 1 "not emp" 0 0
1 2 "emp" 0 0
1 3 "not emp" 1 0
1 4 "emp" 0 1
1 5 "emp" 0 1
1 6 "emp" 0 1
1 7 "not emp" 1 0
1 8 "not emp" 0 0
1 9 "not emp" 0 0
1 10 "not emp" 0 0
1 11 "not emp" 0 0
1 12 "not emp" 0 0
2 1 "not emp" 0 0
2 2 "emp" 0 1
2 3 "emp" 0 1
2 4 "emp" 0 1
2 5 "emp" 0 1
2 6 "emp" 0 1
2 7 "emp" 0 1
2 8 "emp" 0 1
2 9 "emp" 0 1
2 10 "emp" 0 1
2 11 "emp" 0 1
2 12 "emp" 0 1
end
gen spell = long
replace spell = 1 if long[_n-1] ==1 & exit == 1
keep if spell == 1
l, sepby(id)
Scott
----- Original Message -----
From: sara khan <[email protected]>
Date: Wednesday, March 28, 2007 6:16 am
Subject: st: Isolating a single spell
To: [email protected]
> Dear Statlisters
>
> I was wondering if you help me in isolating a single spell from a
> multiple spell series.
>
> I have data on employment status and I wish to extract the longest
> employment spell (indexed by longspl) and also whether the person
> exited employment in the next wave ie exit==1. In the example below,
> for person 1, I wish to keep waves 4-7 as the longest spell ended in
> an exit. For person 2, I wish to keep waves 2-12 even though the
spell
> is censored (no exit took place).
>
> I tried by pid: keep if longspl==1 & exit[_n+1]>=0 but this doesn't
> work. I would be grateful for suggestions.
>
> Many thanks
> Sara
>
>
> +--------------------------------------+
> id wave empstat exit longspl
> --------------------------------------
> 1. 1 1 not emp 0 0
> 2. 1 2 emp 0 0
> 3. 1 3 not emp 1 0
> 4. 1 4 emp 0 1
> 5. 1 5 emp 0 1
> 6. 1 6 emp 0 1
> 7. 1 7 not emp 1 0
> 8. 1 8 not emp 0 0
> 9. 1 9 not emp 0 0
> 10. 1 10 not emp 0 0
> 11. 1 11 not emp 0 0
> 12. 1 12 not emp 0 0
> --------------------------------------
> 13. 2 1 not emp 0 0
> 14. 2 2 emp 0 1
> 15. 2 3 emp 0 1
> 16. 2 4 emp 0 1
> 17. 2 5 emp 0 1
> 18. 2 6 emp 0 1
> 19. 2 7 emp 0 1
> 20. 2 8 emp 0 1
> 21. 2 9 emp 0 1
> 22. 2 10 emp 0 1
> 23. 2 11 emp 0 1
> 24. 2 12 emp 0 1
> +--------------------------------------+
> *
> * 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/
