Little is documented about the association of alcohol consumption and social interaction in Uganda, a country with one of the highest per capita alcohol consumptions in the world. This paper describes the pattern of social interaction by sex and establishes the relationship between social interaction and alcohol consumption with and without the consideration of confounders. The data used had 1479 records and were collected in a survey in 2003. The study was part of a multinational study on Gender, Alcohol, and Culture International Study (GENACIS). Each question on social interaction had been pre-coded in a way that quantified the extent of social interaction. The sum of responses on interaction questions gave a summative score which was used to compute summary indices on social interaction. Principal component analysis (PCA) was used to identify the best combination of variables for a social interaction index. The index was computed by a prediction using a PCA model developed from the selected variables. The index was categorised into quintiles and used in bivariate and multivariate logistic regression analysis of alcohol consumption and social interaction. The stronger the social interaction the more the likelihood of taking alcohol frequently (chi(trend)(2)=4.72, p<0.001). The strength of the association remains significant even after controlling for sex, age group and education level (p=0.008). The strength of relationship between social interaction and heavy consumption of alcohol gets weak in multivariate analysis. Communication messages meant to improve health, well-being and public order need to incorporate dangers of negative influence of social interaction.