The aim of this paper is to assess whether the use of ICT has an impact on student performances as measured in the OECD Programme for International Student Assessment (PISA) 2006. After controlling for observable students’ characteristics and self-selection, we did find a positive and significant effect of the frequency of computer use on science scores. In most countries, however, this effect seems larger when computer is used at home rather than at school. This finding questions the effectiveness of educational policies aimed at promoting computer use at school as a tool for learning.