Creating a column out of the results from a loop

[Lily Ksh]

I have a panel data and I have used a loop to tabulate the standard errors by stacking 3 columns together. The standard error value is estimated for each year (1993-2018).


forvalues i = 1993(1)2018{
preserve
stack column1 column2 column3 if year == `i', into(a) clear
tabstat a, statistics( semean ) columns(variables)
restore
}

The problem is that after estimation, I have to manually input each value into one new column (from 1993-2018). Is there a way to add a command in the loop so that each se value that it estimates is automatically put into the same column?

Thank you.

[Ali Atia]

This may be easier achieved using a reshape rather than a loop:

reshape long column, i(year) j(new)
egen se = sd(column), by(year)
bysort year (new): replace se = se / sqrt(new[_N])
reshape wide

If that doesn't work with your data (which in future posts is best represented by a data example using the dataex command, as explained in the FAQ), try:

gen se=.
forvalues i = 1993(1)2018{
    preserve
    stack column1 column2 column3 if year == `i', into(a) clear
    tabstat a, statistics( semean ) columns(variables) save
    local a`i' = r(StatTotal)[1,1]
    restore
    replace se = `a`i'' if year==`i'
}