January 31, 2018 Girls in the ninth year of elementary school receive better grades in mathematics and Czech language than boys. In anonymously graded tests, however, girls only achieve better results in Czech language, while in mathematics the boys outperform them on average by roughly the same margin.

Meanwhile, the variation in test results within each gender is far greater than the average differences between boys' and girls' results. Gender differences in academic results measured through tests are, in the literature, more often seen to reflect different influences in upbringing and social environments, rather than reflecting different innate dispositions among girls and boys.

When writing non-anonymous school reports for mathematics, teachers give girls better grades than reflect their actual results in anonymously assessed tests. On their reports, girls are given grades for mathematics that are on average 60 % better than their male peers who achieved the same anonymous test score. The gender differences we have found in academic results and grading are in line with the findings of most foreign studies.

Our analysis shows that the differences between boys' and girls' grades is not a result of differences in the way stress affects test results, nor a difference in pupils' liking for or aversion to the subject in question. The likely cause of the difference in grades between girls and boys is that the grades are influenced by the pupils' socio-emotional skills.

Grades on school reports constitute an established means of feedback about each pupil's educational achievements and form one basis upon which aspirations and decisions regarding further study are taken. Biased grades and the incorrect interpretation they may lead to could affect the young people's further academic trajectory in undesirable ways.

Our findings raise a number of questions about suitable changes in how grades are given. One of the many possible changes to be considered is to allocate pupils grades on two separate scales, one of which would reflect only the academic results they have achieved and the second only their attitude to learning, without reference to their objective level of achievement.