Introduction: We examined (a) whether risky drinking behaviour is related to experienced harm from others' drinking (EHFOD) and (b) whether any found relationship is modified by educational level, such that those of lower socio-economic status (SES) experience more harm even when adjusted for drinking behaviour. Method: Data from the Danish national alcohol and drug survey of 2011 (N=5133) were linked with registry data from Statistics Denmark. Eight EHFOD indicators were grouped into nuisance, harassment or harm/damage categories. Indicators for mean alcohol consumption, Alcohol Use Disorder Identification Test (AUDIT) and binge drinking were examined in relation to respondents' EHFOD with multiple logistic regression, stratified by sex and education (proxy for SES). Results: One-year prevalence of EHFOD was 50%. We found a positive and significant relationship between own alcohol consumption and EHFOD categories of harassment as well as harm/damage. Effect modification of education was significant for harassment. Among men, odds ratios for the association between risky drinking behaviour and harassment were 5.50 (95% confidence interval (CI) 3.49-8.65) in the low educational group versus 1.42 (95% CI 0.98-2.07) in the high educational group. Conclusions: Our study confirmed an overall positive relationship between EHFOD and drinking behaviour, but it varied by type of EHFOD. Furthermore, education modified this effect for harassment, suggesting evidence of the alcohol harm paradox with respect to EHFOD. More research is necessary to understand better how drinking patterns diverge between low and high educational groups as well as sex, and how this differentially affects risk for alcohol-related harms, including EHFOD.