§ Mr. Gibb
To ask the Secretary of State for Education and Skills pursuant to his answer of 25 June 2003,Official Report, column 783W, on specialist schools, if he will assess, with reference to sports specialist colleges, the finding in the report by Professc r David Jesson that small differences of up to plus or minus 2.5 percentage points between predicted and actual levels of performance are not statistically significant. 
§ Mr. Miliband
In his report, Professor David Jesson found that 'small differences of up to plus or minus 2.5 percentage points between predicted and actual levels of performance are not statistically significant for schools of average size'. This means that differences of less than 2.5 percentage points are within the uncertainty of the predicted level and therefore should not be heeded at a single school level. Uncertainty in statistical estimates generally decreases with the square of the sample size. For example, increasing a sample 100-fold will generally decrease the uncertainty of estimates from that sample by a factor of 10. Given this, when comparing the difference in performance between groups of schools such as Specialist Sports Colleges with non-selective other schools (95 and 2,342 schools respectively), a difference of plus or minus 2.5 percentage points is statistically significant due to the large number of schools involved.