I am looking for a procedure to construct a current employment/unemployment status variable for people who donot report any employed or unemployed spells in the biography dataset. My intutition is that since these people do not report any new spells in the biography dataset it should mean that their employment status is the same as reported in the last wave. However I am not sure if this intuition is correct. Also, is there a variable that captures this information. I have looked at the variable ts25101(in pTarget) . However I find that his variable changes over waves for a particular ID_t even if these ID_t havent reported any corresponding spells for those wave in the Biography dataset.
the biography dataset is the best starting point to construct a current employment/unemployment status variable. As you probably know, we don`t have information for each participant up to the last wave. Some participants dropped out completely before reaching the last wave, some did not participate in the last wave (but possibly participate in the next wave). So „current employment/unemployment status“ can mean two different things: either the last status of each participant regardless in which wave the participant participated last or the last status of the participants who participated in the last wave. The solution for both problems start by selecting each episode of the biography file where the variable splast „Episode is ongoing“ = 1 „yes“. This will leave all episodes in the dataset that persists until the interview date of the last wave the participant did participate in. If you only want to look at participants who participated in the last published wave you additionally have to select those episode where the wave variable has the value of the last wave (at present this is wave = 10).
There is an additional problem constructing a status variable. A lot of participants hold more than one status at the same time. So if you find more than one ongoing episode for a participant you have to decide which of the status of these episodes should be regarded as the main actual status. For that purpose you have to prioritize the remaining episodes by their status. For example if you regard „employment“ as the most important status, „unemployment“ as the next important status and all other status as the least important, the following STATA code would select the ongoing episode with the highest priority:
keep if splast == 1
gen status = 3
replace status = 1 if sptype == 26
replace status = 2 if sptype == 27
tab status
sort ID_t status
bysort ID_t status: gen primax = _n
keep if primax == 1
By the way the variable ts25101 from pTarget is a bad indicator for the status „unemployment“ because it is only used as a filter variable to collect unemployment episodes under certain conditions. It does not cover each unemployment status and also not systematically actual unemployment status.