Department for Education rapped over use of Sats data
UK Statistics Authority says DfE used data to imply a link between academy status and improvements in test results; plus trust’s land sale, phonics screening and Sats ‘maladministration’.
Does becoming an academy improve pupils’ Sats results? That is the implication of much of the debate coming from the Department for Education, as ministers seek to drive more primary schools, in particular, into academy status.
But is this argument statistically valid? Well, last week the UK Statistics Authority (UKSA) cast serious doubt on it, in a judgment that suggests ministers may have to rethink how they present data on their flagship reform.
The development centres on a DfE statistical publication, released in December and seized on by ministers repeatedly in advocating more academies. This showed, correctly, that Sats results over 2012 to 2014 improved more quickly in sponsored academies – former local authority schools taken over by a sponsor – than in non-academies.
But sponsored academies generally take over schools with poor results, with perhaps more room for improvement than most non-academies, we thought. So was the DfE comparison fair?
In fact, statistical analysis by your correspondent, sent to the statistics authority, suggested that these differences in results gains may not have been anything to do with whether the schools were academies or not. Rather, they seemed to be part of a general trend, for all schools, whereby those starting off with poor results in 2012-13 closed the gap on those that had been formerly been ahead.
And non-academies starting with the same – generally poor – test results as sponsored academies in 2013 actually registered faster improvements in 2014, we found. There was no sign of this analysis in the DfE release.
Read more... (The Guardian - 14 July)