Lagged variables: transition from one condition to another

[Giorgio Piccitto]

Dear users,
I am new of longitudinal analysis and I am running into a problem that may be easy for you:

I need to create one variable which assume values = 1 if the case has passed from the condition of being employed(t-1) to the unemployment(t). All the other transitions have to be coded with 0.

Here an example of my data:

* Example generated by -dataex-. To install: ssc install dataex
clear
input long(pidp apidp) int panel_id byte(wave status)
68028575 68028567 214 1 1
68028575 68028567 214 2 0
68028575 68028571 215 1 1
68028575 68028571 215 2 0
68049655 68049651 317 1 2
68049655 68049651 317 2 0
end

the variable status is: 0 = unemployed
1 = employed
2 = student

I am working with dyads, so each couple parent-son\daughter is identified by a unique id (panel_id)

Do you have any clue about that?

thanks a lot, G

[Giorgio Piccitto]

I think I solved with
gen transi = ( status == 0 & L.status==1 ) ///
if !missing(L.status)

sorry for this post, but maybe someone will find it useful.

G

[Clyde Schechter]

Never apologize for posting. The probability is very high that somebody else will have a very similar question in the future and will come here searching for an answer. You've made it possible for them to find it. Thank you!!!

[Nick Cox]

Notice that this simplifies to

gen transi = status == 0 & L.status == 1

as if the previous value is 1 then it is necessarily not missing too.