In this article we evaluate wage differentials in Italy combining gender and education perspectives. The main goal of this article is to verify whether the extent of the gender pay gap varies between highly- and low-educated workers, and whether or not the role played by gender differences in characteristics and in market rewards is similar in the two groups. We apply quantile regression analysis and an adaptation of the procedure suggested by Machado and Mata (2005) to evaluate the predicted wage gap at different points of the female wage distribution scale. The analysis is carried out on the Italian sample of the last available year of the European Community Household Panel (ECHP). We show that the extent and the trend of the gap predicted across the female distribution is sharply different between groups with diverse educational levels. In the case of low-educated workers, although the predicted gap is largely explained by differences in rewards, lower levels of education or experience are responsible for the gap, especially on the right-hand side of the distribution. On the contrary, highly-educated females have better characteristics than highly-educated men that partially compensate the rather high difference in returns, in particular at the extremes of the distribution.