This study presents a quantitative approach based on Differential Item Functioning analysis within the Item Response Theory framework, to quantify gender differences in tackling specific mathematical items. We use this approach to explore two crucial topics in mathematics education (misconceptions in decimal numbers, and estimation) by analysing answers given by Italian students to specific mathematical items taken from a sample of 1400 items administered in INVALSI tests since 2008. For each item, we have a sample of 30,000 students per year, statistically representative of the entire Italian student population. The results section presents a didactic interpretation of the statistical evidence and shows how interdisciplinarity between statistics and mathematics education, with a mixed-method approach (Johnson & Onwuegbuzie, 2004), represents a good approach to exploring the gender gap in relation to specific constructs of math education. This approach, common in many disciplines, can also make an interesting contribution to mathematics education research.