Accuracy describes the closeness between an estimated result and the (unknown) true value.
Data coverage
The target population for the phonics data collection is all year 1 pupils in state-funded schools and all year 2 pupils in state-funded schools who had not passed the phonics screening check in year 1 (including those who did not take it in year 1). The data may not be representative of all 6 year old pupils in England since those in independent schools and those who are home-schooled will be excluded.
LAs can submit further data (either additional data or amendments to that already submitted) after the initial deadline of the collection until mid-September. The collection is then closed and no further changes can be made.
There is very little change between the provisional and final datasets and the statistical release is produced based on the provisional data.
Any pupils who do not have a valid phonics outcome are excluded from the calculations and do not appear in the number of eligible pupils or in the outcome percentages. Valid results in the phonics screening check are:
| Phonics outcome code | Description |
|---|---|
| Wa | Took the phonics screening check and met the required standard |
| Wt | Took the phonics screening check and did not meet the required standard |
| A | Absent |
| D | Did not take the phonics screening check |
| Q | Maladministration of the phonics screening check |
Pupils with 'D' are included as we want to measure the percentage of all 6 year olds meeting the standard rather than of only those who took the check.
Steps are taken to minimise the number of pupils who are absent (for example those absent during test week can take the check the following week). Those who are absent are also included in calculations to encourage schools to ensure that all eligible pupils take the check where possible.
Pupils whose phonics screening check outcome was subject to maladministration are also included to ensure complete coverage of the cohort.
In addition to figures on the percentage of the cohort who pass the phonics screening check in year 1, we also publish the percentage of pupils who pass by the end of year 2. We do not publish figures for the percentage of those who take it in year 2 who pass. See annex C for a precise definition of the year 2 cohort.
Measurement error
Measurement error is the difference between the actual value of a quantity and the value obtained by a measurement. Repeating the measurement will reduce the random error caused by the accuracy of the measuring instrument but not any systemic error caused by incorrect calibration of the measuring instrument.
The following steps are taken to minimise measurement error in the assessments.
Clear guidance (opens in a new tab) (opens in a new tab) is provided to schools to ensure that the phonics screening check is administered consistently in all schools. Since 2014, the pass mark has not been made available to schools until after the completion of the check (in previous years, it was sent out with the test materials). This was to ensure that teachers could not be influenced by the pass mark when judging whether pupils had read each word correctly.
In addition, LAs carry out monitoring visits to at least 10% schools in their area, before, during and after the check period to ensure that they are following the security arrangements and administering the phonics screening check in accordance with the published guidance. They will inform STA of any irregularities in schools’ assessment arrangements.
STA will investigate any matter brought to its attention relating to the accuracy or correctness of any child’s check results. This can lead to changes to, or annulment of, results for a whole cohort, groups of children or individual children.
The phonics collection remains open for six weeks after the initial collection deadline so that revised data can be submitted to correct any errors identified by schools or LAs after submission.
Validation and quality assurance of source data
Once schools have finalised their phonics marks , they must enter them into their management information system (MIS) which will already contain details of all pupils in each year group. The MIS will include a number of validation rules which check that the data entered is valid and alert the school to correct the data if not.
Once all results have been entered, the school will instruct the MIS to generate a standard file that they will transfer securely to the LA. The LA will check the data before forwarding it to the department using COLLECT. COLLECT also contains validations rules (similar to those in school MIS) which highlight ‘errors’ and ‘queries’ to the LA.
The LA must correct any errors before they can approve the data. ‘Queries’ highlight situations which are not usual but can occasionally occur (for example, the number of boys submitted by a school is more than 10 fewer than the expected number). Before the LA can approve the data, they must enter a note against these items to confirm that the data is correct.
| Examples of validation rules |
|---|
| The phonics mark must be a number between 0 and 40 |
| Pupils with an outcome of A (absent) or D (did not take) must not have a phonics mark |
| Each pupil must have a unique pupil number (UPN) in the correct format |
The LA must also ensure that phonics data is collected from every expected school.
The department operates a data collection helpline to assist LAs that are having difficulty submitting data. The helpline will also contact any LAs who they believe may be having problems submitting their data or issue a guidance note to all LAs if the same errors are repeated in many schools’ data.
The department monitors the level of returns and the helpdesk contacts LAs with outstanding data as the submission deadline approaches.
If the volume of data submitted is below expected levels when we are due to take the cut of provisional data, we would analyse the data to determine whether it was sufficiently representative nationally and at LA level. Decisions would then be taken whether to extend the collection for a day or two (or longer if necessary though this would probably delay publication of the statistical publication) or to suppress data for any LAs where coverage was felt to be a particular issue.
Occasionally, more than one school may submit data for the same pupil (for example, if pupils change school during the summer term). Most duplicate cases are resolved when collected data is matched into the National Pupil Database.
Data processing
Within the national pupil database (NPD), data on pupil’s attainment from the phonics data collection is linked with information on pupil’s characteristics taken from the school census (opens in a new tab) (opens in a new tab). To enable this linking, records are matched, using fields such as surname, forename, date of birth, UPN, gender and postcode. This successfully matches around 60 to 75% of pupils. Additional, more complex, routines are then applied to match as many of the remaining pupils as possible, up to around 98%.
Occasionally, a pupil will appear more than once in data, resulting, for example, from a change of school, or dual registration. Rules for deriving the main record and a combined ‘best’ attainment record for these pupils have been agreed. Where a pupil has more than one result in a subject, the highest level will be taken and all other results discounted. Occasionally a pupil will appear more than once on the census. Rules for deriving the main census record have been agreed to eliminate duplicates based on factors such as enrolment and school type.
Statistical publication production
All data in the publication tables are produced by one person and quality checked by another. Any discrepancies in the data produced are discussed and more experienced staff involved as required to agree the correct figures. Additional checks are also carried out on the data produced.
Examples of additional checks:
- Comparisons with previous figures to identify any large changes
- Check totals are consistent across tables
- Check patterns in the data are as expected
Disclosure Control
The Code of Practice for Official Statistics requires us to take reasonable steps to ensure that our published or disseminated statistics protect confidentiality.
Figures for the Isles of Scilly and City of London are suppressed in the phonics publication as these LAs have a single school and we do not publish school level information for phonics. Regional eligible pupil figures are rounded to the nearest 10 so that it is not possible to derive figures for these LAs by summing the figures for the other LAs in the region.
Reliability
Reliability isthe extent to which an estimate changes over different versions of the same data.
The phonics publication is published based on provisional data as there is very little change between the provisional and final datasets (the only change being the submission of a small amount of late data). The national figures typically show no change between provisional and final data. Table 1 shows the change in some key figures at national level over time.
Table 1: Change in national phonics data
| 2012 | 2013 | 2014 | 2015 | 2016 | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| % of pupils meeting the expected standard of phonics decoding by end of year 1 | Provisional Final | 58% 58% | 69% 69% | 74% 74% | 77% 77% | 81% 81% | 81% 81% | 82% 82% | 82% 82% | No assessments | 75% 75% | 79% 79% | 80% - | |
| % of pupils meeting the expected standard of phonics decoding by end of year 2 | Provisional Final | - - | 85% 85% | 88% 89% | 90% 90% | 91% 91% | 92% 92% | 92% 92% | 91% 91% | 87% 87% | 89% 89% | 89% - | ||
Phonics mark distribution
In 2024, the mean mark was 33 and the median was 36. Any change in the percentage of pupils achieving each mark is influenced by changes in the difficulty of the check, as well as the ability of the cohort. The standard is anchored only at the expected standard (32 marks), therefore the standard required to reach other marks is not exactly equivalent year-on-year.
