I remember reading it first in Reason.

The problem was that the longitudinal study created to see whether D.A.R.E. would work was still in its midst when all the anti-drug crusaders got jealous that only some students had this drug-abuse-prevention program. So it spread far beyond the test subjects and got widely adopted before the data came back that D.A.R.E. students were a little more likely to try illegal drugs than those without D.A.R.E. ( albeit within the margin of error. )

And it still shows no sign of going away despite being abandoned by its creators as a failure.

Dante