Education settings survey
Non-response adjustments made to published figures from 9 September 2020 are summarised here. For the non-response adjustment methodology made to published figures from 23 March to 16 July 2020, see the July 21 publication. Non-response adjustments have not been made to local authority level data, see the ‘Attendance by local authority’ section for more information.
The response rate among state-funded schools was 84% on 11 February. Open rate and attendance figures for state-funded schools in this release are adjusted to account for those that did not respond to the survey.
Methodology in 2020/21 autumn term
The methodology used from 9 September assumes that state-funded schools that did not respond were no more likely to be closed, ask pupils to self-isolate or have lower attendance than responding state-funded schools. This follows analysis of response patterns, data collected from a sample of non-responding schools, and comparison with data collected through regular phone calls between DfE and local authorities to identify any closed schools. Figures for state-funded schools have been weighted to take account of differences in response rate between different school types.
To reflect the number of local authorities that were on half-term, adjustments were made to the methodology for figures between 19 and 23 October and 2 November.
Response rates among state-funded schools fell on 17 December by 7 percentage points compared to the previous day. Some schools reported inset days or non-COVID related closures on 17 December (1% of responders). Schools are not required to complete the form if on a planned holiday, such as Christmas break and, previously, decreases in response rates have been associated with an increase in school closures. As a result, it is likely there is a greater proportion of closures among non-responders than responder and we do not know the distribution of these between i) COVID-related closures and ii) schools starting Christmas holidays or having inset days as planned. Therefore, estimates for 17 December are less reliable than previous days and likely over-estimate open and attendance rates. Comparisons within the commentary here made here with Wednesday 16 December instead. Data for 17 December is still available in the underlying data.
Methodology in 2020/21 spring term
The methodology used from 11 January assumes that state-funded schools that did not respond were no more likely to be closed to vulnerable children and critical worker children or have lower attendance than non-responding schools.
This follows analysis of response patterns and data collected from a sample of non-responding schools to identify any closed schools. This found that up to 10% of non-responding state-funded primary schools and state-funded special schools were closed, compared to less than 1% and around 1% of responding schools respectively. Adjusting for this would reduce open rates for these school types by up to 1 percentage points, but would have a negligible impact on attendance rates (less than 0.5%) given high response rates (over 80%) and relatively low attendance rates. Therefore, no adjustments have been made at this time but we will continue to review the non-response methodology. Figures for state-funded schools have been weighted to take account of differences in response rate between different school types.
Response rates for independent schools are lower than other settings (less than 60%) and, following review of a sample of non-responding independent schools, we found non-responding independent schools were more likely to be closed than responders. Figures for independent schools have not been adjusted to account for this response bias and are presented based on responding schools only.
Local authority early years survey
For the local authority early years survey, figures for the number of settings open and closed and the number of children attending for those LAs which did not respond are estimated based on the proportions reported by LAs which did respond. The national estimate comprises reported figures from LAs which did respond combined with these estimates (see the methodology section). From 1 June onwards, this methodology has been revised to take into account the wider opening of early years settings.