The number of states assigned to treatment is randomly determined, but at least 25% of statesmust be classified as treated and at least 25% of states must be classified as control in anypseudo-experiment.  

我只会一半一半
local STATES = 51
local PERIODS = 120

gen t = runiform() if period  == 1

egen rank = rank(t)
gen T = .
        replace T = 0 if rank >= 1 & rank<= `STATES'/2
        replace T =1 if rank >`STATES'/2 & rank<= `STATES'


如何才能 随机25% 呢？