I have an unbalanced panel of 7 years with every person interviewed 4 times and I want to drop all the people that reported that they were unemployed/inactive in all 4 periods. However, I do not want to drop the observations of the people that may have been out of the labour market for 1, 2 or 3 out of the 4 periods they were interviewed. How do I tell Stata to drop people based on their situation in multiple years (t to t-3)? When I do drop if ecostatus>3
, for example, Stata drops observations that I need, i.e. the people that were inactive for less than the full period of the survey.
// create some example data
clear
input id t unemp
1 1 1
1 2 1
1 3 1
1 4 1
2 1 1
2 2 0
2 3 1
2 4 1
end
// create the total number of unemployment spells
bys id : egen totunemp = total(unemp)
// display the data
sort id t
list, sepby(id)
// keep those observations with at least one
// employment spell
keep if totunemp < 4
// display the data
list