Controlled drinking has long been a controversial topic in behavior th
erapy. The historical context of this debate is reviewed, with special
attention paid to the pioneering research conducted by Mark and Linda
Sobell (the first behavior therapists to publish a controlled trial o
f controlled drinking with alcoholics). After updating the research fi
ndings and predictors of controlled drinking with alcohol-dependent dr
inkers, literature on the effects of moderation training (including br
ief interventions) designed to reduce the risks of alcohol abuse is re
viewed. As an illustration of this approach, preliminary data are pres
ented from an ongoing study investigating the effects of a stepped-car
e secondary prevention program for high-risk adolescent and young-adul
t drinkers. Throughout the paper, harm reduction is presented as an ov
erarching model of behavior change that encompasses both controlled dr
inking for alcohol dependence and moderation training in the preventio
n and treatment of alcohol abuse. Unlike abstinence-only or ''zero-tol
erance'' approaches, the harm-reduction model supports any behavior ch
ange, from moderation to abstinence, that reduces the harm of problems
due to alcohol.