All large scale surveys on reading show that girls have on average significantly higher test results than boys. Data from large scale surveys have been used to further analyze the gender difference in reading. Boys are overrepresented among the weakest readers. In the Swedish parts of the PISA studies in 2000, 2003 and 2006 the proportion of boys in the two groups at the lowest level of reading competence has been twice as high as the girls proportion. In the Swedish part of PIRLS 2001 ten percent more boys could be found among the lowest performing students and ten percent more girls among the highest performing students. The same pattern was found in PIRLS 2006. Ryan och DeMark (2002) raised the question whether gender differences in reading were related to the type of items presented in tests and the subject of the items. In an analyze of data from IEA 1991 Taube and Munck (1996) looked at the gender difference in reading to see whether it was related to the content of different texts, the cognitive demands in the items, the length of the texts and the number of task linked to the items. An “inequlity index” was used to measure the proportion of boys and girls who had responded correctly to the different items considering the difficulty of the item. The results showed that items about human beings, animals behaving as human beings and human activities were more favourable for girls than items about neutral subjects or science oriented subjects. This could be seen in all participating countries. It was also shown that girls performed better in relation narrative texts and that long text favoured girls more than short texts. Among the fourteen year old students there was also a difference between boys and girls related to whether the main character of the text was male or female. Boys performed significantly better when reading texts with males as main actors. For the girls the gender of the main character did not seem to influence the results. Such differences were not found among the nine year old students. These findings are in line with earlier findings in a study by Asher and Markell (1974). They found that boys were more sensitive towards the content of the text than girls. Boys performed better when they found the texts interesting, while girls performed at the same level independently of whether they found the text interesting or not. A study by Scott (1986) indicated that boys seemed to perform better when the content of the text seemed to be more ”suitable” for boys than when it did not seem to be. Data from the PISA studies have been used for similar analyzes. Roe and Taube (2003) found that the largest difference in girls favour where when the students were asked to respond to items containing continuous texts, especially narrative, argumentative and injunctive texts. The difference between boys and girls was smaller when the item contained non-continuous texts such as charts, maps and diagrams.

Mid Sweden University and Kalmar University have organised surveys on reading skills in Swedish municipalities. This paper will focus on the surveys made in four municipalities between 2002 and 2007. All students in grade 8 in these four municipalities have participated in the surveys, in total 16 287 students. These surveys have included the first part of the reading test used in the IEA: Reading Literacy study in 1991 and information collected from the teachers about the students, including gender. The results from the reading test will be analyzed in order to measure the proportion of boys and girls who have responded correctly to the different items considering the difficulty of the item. The ways boys and girls have responded will be analyzed in the perspective of items containing different types of texts, short and long texts and different characters appearing in the texts.

The results of the analysis are expected to be in line with similar analysis made in 1996 of all students (Swedish students as well as students from other countries) in the IEA Reading Literacy Study in 1991 and in 2003 of the Nordic students participating in PISA 2000. It is expected that girls will do much better than boys on longer items, continuous texts and texts about human beings. Boys on the other hand are expected to narrow the gap to the girls on shorter items, non-continuous texts and texts on more neutral subjects. Of special interest is to see how similar the patterns among the students in grade 8 in these four municipalities in 2002 – 2007 are with the patterns among Nordic students in grade 9 in 2000 and students from all parts of the world in 1991.