Data collected from the alcohol prevention program «Adolescents and Alcohol» [Ungdom og Alkohol] among Norwegian 8th grade students, was analyzed to assess implementation quality. Only data from the intervention group; 586 students, was utilized, to see if aspects of the implementation could predict positive program outcomes reflected in changes in self-reported alcohol consumption. As many students yet have to try alcohol for the first time at age 14, alcohol expectancies are already established, and predict alcohol use. Alcohol expectancies were therefore used as an alternative measure to alcohol use. Durlak and DuPre`s (2008) framework for measuring implementation quality was used, and student self-reports reflect implementation characteristics presented in this framework. Results from analyzing individual scores indicate that a certain amount of teacher control is positive, as well as seeing that the whole video project intervention was delivered in full. These results indicate that fidelity and dosage are aspects of the implementation which have an influence on alcohol use in this study. Results differ when scores are aggregated to represent whole classes instead of individuals; classwise meanscores. More student involvement predicts decrease in both alcohol use and alcohol expectancies, and these results reflect quality of program deliverance and fidelity. No gender differences in alcohol use were reported, but differential alcohol expectancy scores between T1 and T2, showed a larger increase in boys than girls. The conclusions point to several important aspects with implementing the Adolescents and Alcohol prevention program, regarding aspects of program delivery and staff characteristics. The conclusions support earlier implementation literature, which states that implementation influences program outcome, and that steps should be taken to ensure that implementation is assessed and verified.