Hi :wave:
I'm having trouble with the data management necessary to set the database so that I can run a Lifetable to do some descriptive analysis.
I tried a simple snapspan synthax:
snapspan id year x1-x15, gen (year0)
but i'm getting an error message (which said "not sorted", even if I sort the data for id or id and year and put the variables x1-x15 in the order that they have in the database) and a theorethical issue. Stata says that it will ignore a big quantity of observations, because I've just one record for those id, but it happens because some people drop out from University during their freshman year. What can I do not to lose that piece of information? I found on statalist a solution to this problem and I tried it, but it didn't work. Firstly, because it used expand with by, and secondly (after i fixed that thing) stata said to me that it didn't make any change :shakehead
this is the sinthax (the second line is the one I modified, and I've tested it until that)
sort id
expandcl 2 if _N==1, gen (expanded) cluster (id)
replace year = . in <originalobs>/l
sort id year
by id: replace year = year+1 if year==1 & _n==2
The variable year contains integers from 1 to 5 (first, second year of university and so on). It is repeated if you need an extrayear to finish (Italian University System, speaking of which, sorry for my English!
), then there are two 3 for the same id if the person needed one extrayear for his bachelor degree.
Thanks for your help!
best regards ^_^
I'm having trouble with the data management necessary to set the database so that I can run a Lifetable to do some descriptive analysis.
I tried a simple snapspan synthax:
snapspan id year x1-x15, gen (year0)
but i'm getting an error message (which said "not sorted", even if I sort the data for id or id and year and put the variables x1-x15 in the order that they have in the database) and a theorethical issue. Stata says that it will ignore a big quantity of observations, because I've just one record for those id, but it happens because some people drop out from University during their freshman year. What can I do not to lose that piece of information? I found on statalist a solution to this problem and I tried it, but it didn't work. Firstly, because it used expand with by, and secondly (after i fixed that thing) stata said to me that it didn't make any change :shakehead
this is the sinthax (the second line is the one I modified, and I've tested it until that)
sort id
expandcl 2 if _N==1, gen (expanded) cluster (id)
replace year = . in <originalobs>/l
sort id year
by id: replace year = year+1 if year==1 & _n==2
The variable year contains integers from 1 to 5 (first, second year of university and so on). It is repeated if you need an extrayear to finish (Italian University System, speaking of which, sorry for my English!
), then there are two 3 for the same id if the person needed one extrayear for his bachelor degree.Thanks for your help!
best regards ^_^