Although the gender differences in TIMSS-2003 in grade 4 were smaller compared to TIMSS-1995, boys significant outperformed girls in mathematical achievement. Furthermore, the decrease in gender differences was not the result of an improvement in the mathematical skills of girls. In the last eighth years, the overall average mathematics achievement of Dutch grade 4 students decreased. Further analysis showed that this decline was caused by the decrease in average achievement of boys; the achievement of girls did not change. The international report showed that the Netherlands was the only country in which this occurred (Mullis, et. al, 2004). In other countries, improvements or declines in achievement were consistent between boys and girls. The attitudes of grade 4 students towards mathematics was indifferent; they did not like or dislike mathematics. However, grade 4 boys enjoyed mathematics significant more and showed a higher confidence in their mathematical skills than girls (Meelissen & Doornekamp, 2004a; 2004b). Gender differences in both mathematics achievement and attitude in grade 4 are the main focus of this study. This paper explores the influence of school, class and student characteristics on the mathematics achievement of girls and boys. The main research question is: "To what extent can school, class and student characteristics, and especially students' attitude towards mathematics, explain variance in mathematics achievement of girls and boys in grade 4?" The analyses are carried out with the multilevel programme MlwiN, which takes the nested design of this study (within each school the students of one grade 4 class were tested) into account. For the final analyses, the data set is split: a data set for girls and a data set for boys. The results of these analyses will provide more insight in (non-changeable and changeable) factors influencing the mathematics achievement of girls and the mathematics achievement of boys.