Purpose: The purpose of this paper is to describe and analyse the implementation by a non-governmental organization (NGO) of an intervention with two different strategies--one employing confrontational approaches, the other cooperative ones--aiming to reduce the rate of successful purchase attempts (PAs) of medium-strength beer in Sweden. Design/methodology/approach: The Swedish Youth Temperance Movement (UNF) has been responsible for the development and implementation of the intervention, and a research team at Orebro University for the evaluation. The outcome analysis is based on 1,475 PA from 25 Swedish cities during 2003-2006. Annual reports, discussions and telephone interviews have been used to understand the intervention process. Findings: In 40 per cent of all registered PAs, beer is sold to minors. In the eight cities using a structured strategy, there is a significant decrease from 2003 to 2006 (44-27 per cent; p less than 0.001). The confrontation method, compared to no model, is more than four times (OR = 3.8; CI 2.0-7.0) more likely to yield a positive result. Compared to the cooperation method, it is even more likely to yield a positive result (OR?= 4.7; CI 2.1-10.7). Practical implications: Working with a structured strategy gives significantly better results than working without one. The confrontation method is more successful than the cooperation method. Developing a preventive strategy takes time, cooperation perhaps requiring even more time to succeed than confrontation. The result reveals the importance of building local networks in the municipalities. Originality/value: The advocacy for Swedish alcohol policy by the young members of UNF and participatory research is a unique combination. (Contains 6 tables and 4 figures.)