The code below shows how to create an id variable that takes on the values 1
to n within each group/panel (where n is the number of cases in a
group). The grouping variable is defined by the user. The number of cases, n, need not be
the same across all groups. In the example dataset, group is the grouping
variable. The variables v1 and v2 are additional variables in the dataset
and are not modified by
the code.
data list list /group v1 v2.
begin data
1 12 0
1 23 1
1 22 0
3 5 0
3 7 0
3 13 1
5 1 0
5 56 1
end data.
sort cases by group.
compute id = 1.
if group = lag(group) id = lag(id) + 1.
exe.
list group id v1 v2.
group id v1 v2
1.00 1.00 12.00 .00
1.00 2.00 23.00 1.00
1.00 3.00 22.00 .00
3.00 1.00 5.00 .00
3.00 2.00 7.00 .00
3.00 3.00 13.00 1.00
5.00 1.00 1.00 .00
5.00 2.00 56.00 1.00
Number of cases read: 8 Number of cases listed: 8