New research from the University of Warwick suggests girls benefit significantly from more interaction with very bright peers at secondary school, but it can be detrimental for boys.
In a paper published in the latest edition of the Journal of Labor Economics, a research team from the University of Warwicks Economics department and the London School of Economics looked at data from all secondary schools in England to determine the effect of ability of peers.
They found that a large number of bad peers in classes, defined as children in the bottom 5% of the ability range, has a significant negative effect on other classmates. However, a large fraction of good peers, defined as children in the top 5% of the ability range, appeared to benefit girls, but had a detrimental effect on boys.
The researchers used results from national tests taken at ages 11 (Key Stage 2) and 14 (Key Stage 3) and estimated both the effect of average peer quality and the effect of being at school with a high proportion of very low-ability and very high-ability pupils.
Lead author on the study, Professor Victor Lavy from the University of Warwick said: Our results imply that a 10% decrease in the numbers of bad peers at school is associated with a 10% improvement compared to the average for Key Stage 3.
However, a 10% increase in the numbers of good peers in a class is associated with a 10% increase in performance for girls only, compared to the average for Key Stage 3. Boys actually seem to lose about 5% in performance.
Professor Lavy said although the differences between the results for boys and girls were surprising, it was impossible to say if they lent support to the idea of tracking students by ability.
He said: If for example, we divide the class into two ability groups, boys and girls in the lower group will experience a significant decline in academic performance while in the higher ability group girls will benefit and improve their learning while boys will not be affected.
So do our results lend overall support to tracking of students by ability? There is no simple answer to this question, but we believe our data is rich enough and our findings robust enough to provide a solid ground for insightful interventions targeting students ability mix as a means to improve learning standards.
Explore further: Scientists seen as competent but not trusted by Americans