Clinical observations suggest that girls suspected to have ASD manage to perform better in social and emotional tasks than boys, leading to a camouflage effect which results in the under-diagnosis of girls with ASD. We used the Social Attribution Task (SAT) in order to assess the performance of 12 girls who were previously diagnosed with ASD, and additionally to compare their performance in SAT to that of 12 boys who were previously diagnosed with ASD. No significant differences were found between girls and boys in all the indices of SAT except for the Person index; girls performed better than the boys in the Person index. The lack of significant differences between the boys and the girls in five out of the six SAT indices suggests that girls with ASD do not have a better social understanding than boys. They outperformed the boys only when the instructions included a clue that triggered them to name human traits, as they are instructed to do in the Person index. Implications of the usefulness of the Social Attribution Task (SAT) for diagnosing ASD in girls are discussed.