1920 QAR error

Ah, but it implies it's only on the overall calculation for the NARTs, rather than our individual provider level stats (which, of course, weren't published), which would explain why we didn't spot it.

Nick Linford and I discussing it on twitter. My first thought was they'd used R12, Nick reckons they didn't include R04 (although I've never quite understood what they take from R04 and what gets counted as the next year).

I suspect it is due to late notification by Providers due to Covid that now shows the QAR was overstated.

But Martin, that happens every year (which is why late (post-R04) reported withdrawals count against you in the next year)? This is implying they didn't use the stated dataset correctly, per my suggestions above.

I'm with you Steve - I suspected an R04 error but a difference with the national calculation seems more likely as providers would have picked this up (which is one of the reasons we pushed back on them not publishing R10, 12, 13 versions in that year but without success)

https://explore-education-statistics.service.gov.uk/find-statistics/apprenticeships-and-traineeships

Error in apprenticeship QARs for 2019/20 (as published in March 2021)

QARs are produced each year using a multi-year pooled data set and are subsequently published as official statistics.  A calculation error has been discovered which affects the 2019/20 national level rate which was originally published in March 2021.  The error has identified people who had withdrawn from their apprenticeship programme but who, at the time of publishing, were thought to have continued into the next year.

In line with the code of practice for statistics we have taken down rates for 2019/20 pending corrected data being available.  We apologise for the error and for this data not being available.  This is necessary to avoid any further use of the numbers which we know to be incorrect.

We need to complete the full production cycle for 2020/21 in order to provide new data for 2020/21 and revisions to any preceding years which will include full quality assurance of the 2019/20 data.  The error is likely to lead to a reduction to the overall rate that was previously published, but the size of the impact can only be robustly quantified once the revised data is calculated.  This revised data for 2019/20 will form part of the national high level three-year time series, containing data for 2018/19 to 2020/21, which is scheduled for publication in March 2022.  Once that data has been calculated and quality assured, national level data for 2019/20 will be available again.  This is scheduled to be published at the end of March 2022.

Education and Training achievement rate statistics published in March 2021 are unaffected by the error and are available here.