Tables 'can never show value added'
League tables are fundamentally flawed and the Government is wrong to think they can be reformed to show the value added by schools, a leading educational statistician has warned ministers.
Performance tables cannot give an accurate picture because they are too simplistic, Harvey Goldstein, professor of statistical methods at the Institute of Education in London, said this week.
But policymakers had allowed themselves to be blinded by science into supporting what are really "alarmingly simple models" disguised by the complex calculations needed simply to obtain decent-looking results at the end, Professor Goldstein said.
Tables were among the most contentious of the previous government's education reforms. But Labour wants to extend their use - including for the first time publishing performance data on teacher-training providers.
A consultation paper was published by the Government in March on potential reforms to GCSE tables to show value added.
But Professor Goldstein, delivering a professional lecture at the institute on Wednesday, said repeated studies had shown that even with value-added weighting, valid comparisons between most schools would still be impossible to make.
"It is this finding above all that provides strong evidence against current policy on the publication of league tables - of whatever kind," he said.
Indeed, the Government's consultation paper commended what he called a particularly crude form of value added, precisely because it did not depend on a complex mathematical model. "There is no justification in using a procedure because it is simple."
Professor Goldstein said that schools were complex systems, with pupils' performance influenced by peer groups inside and outside, by home life, and by previous schools attended, "If we wish to capture such complex relationships, then we need to have tools which are capable of matching that complexity, " he said.
A study of results in London schools published last year by Professor Goldstein and colleague Pam Sammons showed that junior schools had three times the influence over pupils' GCSE results that secondary schools had. Secondary schools accounted for only 2 per cent in the variation in GCSE results between pupils.
There is an irony in the current debate over the use of league tables. Professor Goldstein said it echoed the debate of 20 years ago about whether educational standards had fallen. A Government-appointed unit later concluded "reluctantly" that it was impossible to measure trends over time.
With the Government now setting literacy and numeracy targets for 11-year-olds, Professor Goldstein said: "One can only hope that those politicians willing to stake their careers on such notions are familiar with recent history."
* Letters, page 20