Aims: To estimate (1) the associations between parent-adolescent relationship, parental knowledge and subsequent adolescent drinking frequency and (2) the influence of alcohol use on parental knowledge. Design: Path analysis of school based cohort study with annual surveys. Setting: Post-primary schools from urban and intermediate/rural areas in Northern Ireland. Participants: A total of 4937 post-primary school students aged approximately 11 years in 2000 followed until approximately age 16 years in 2005. Measurements: Pupil-reported measures of: frequency of alcohol use; parent-child relationship quality; subdimensions of parental monitoring: parental control, parental solicitation, child disclosure and child secrecy. Findings: Higher levels of parental control [ordinal logistic odds ratio (OR) = 0.86, 95% confidence interval (CI) = 0.78, 0.95] and lower levels of child secrecy (OR = 0.83, 95% CI = 0.75, 0.92) were associated subsequently with less frequent alcohol use. Parental solicitation and parent-child relationship quality were not associated with drinking frequency. Weekly alcohol drinking was associated with higher subsequent secrecy (beta -0.42, 95% CI = -0.53, -0.32) and lower parental control (beta -0.15, 95% CI = -0.26, -0.04). Secrecy was more strongly predictive of alcohol use at younger compared with older ages (P = 0.02), and alcohol use was associated less strongly with parental control among families with poorer relationships (P = 0.04). Conclusions: Adolescent alcohol use appears to increase as parental control decreases and child secrecy increases. Greater parental control is associated with less frequent adolescent drinking subsequently, while parent-child attachment and parental solicitation have little influence on alcohol use.