I read a lot of alcohol intervention studies, and most look for the fanciest add-on: gamification, AI coaching, peer support. This one did the opposite. It stripped an intervention down to almost nothing - just asking people to notice their own drinking - and it still worked.

What they did

Researchers in Korea recruited 94 college students who'd already flagged as hazardous drinkers on a screening survey. In year one, everyone did the same thing: 14 days of smartphone surveys, four times a day, logging how much they drank, their mood, hangovers, and any negative consequences. That's ecological momentary assessment (EMA), real-time self-tracking instead of a single retrospective questionnaire.

A year later, the group split. 49 students repeated the same 14-day tracking with nothing added. The other 45 got a personalized feedback report before the second round started, covering their drinking patterns from year one, their emotional states, and even how their genes process alcohol, plus daily feedback messages throughout the second tracking period.

What they found

  • Both groups drank less afterward. Hazardous drinking scores, measured with the AUDIT-C screening tool, dropped modestly but significantly right after the tracking period, in both conditions.
  • The drop held up. At the three-month follow-up, neither group had slid back.
  • Drinking motives shifted too, but later. Social and "enhancement" motives (drinking to feel good, not just to fit in) were lower at three months, though not right after the intervention.
  • Expectancies didn't budge. What students believed alcohol would do for them stayed the same throughout.
  • The feedback group did no better. Genetics, mood data, personalized coaching messages, none of it beat simply tracking with no feedback at all.

What it means

The authors' honest conclusion: they can't explain why the fancier version didn't help more. My best guess, reading between the lines, is that the tracking itself was already doing the work. Answering "what did you drink, how do you feel, any consequences" four times a day for two weeks forces a kind of attention that a report handed to you afterward can't replicate. You're not being told something about yourself. You're catching yourself in the act, repeatedly, until the pattern is hard to ignore.

Worth flagging that there's no true do-nothing control group here, so part of the improvement could be students settling into a new school year rather than the tracking itself, but the effect held for three months, which argues against pure noise.

That's basically the same principle I built AlcoBalance around. It doesn't hand you a verdict on your drinking after the fact. It shows you your pace and your peak while the night is still happening, so you're the one noticing and deciding what to do about it, not reading someone else's report the next morning.

Source: Addictive Behaviors, DOI