International standardized assessments highlight the existence of gender differences in mathematics performances: male outperform female in many countries and in Italy this gap is remarkable. The results of Italian standardized assessment, called INVALSI test, confirm the existence and the importance of this gap. In this paper we examine three INVALSI items for Grade 10, similar in terms of mathematical content, item type and question intent. The quantitative analysis of each item, based on the Rasch Model, shows that male and female answers have distinctly different response behaviors, and these differences are similar in the three items. A pedagogical interpretation of the quantitative results led us to explain this gap in terms of a different influence of a specific misconception on males and females.
Keywords: Gender gap; misconceptions; percentages; Rasch analysis; standardized assessment.