The statutory returns

Regulatory returns are a reading of the record, not a project.

Kestrel produces the census, the Address Collection the government uses to estimate capacity to contribute, STATS attendance and the NCCD counts, from the record you have been keeping all year. Each stays defensible eighteen months later when somebody asks where a number came from.

Watch the walkthrough All modules

At a glance

Kestrel produces the non-government schools census, the Address Collection, STATS attendance and NCCD from the student record, with the evidence behind each figure.

Runs against a moment

Census, NCCD, Address Collection and STATS read the record as at a moment.

Gaps named, by student

A missing value stops the run and names the student and the field.

The substantiation pack

The rows behind each person’s figures, as a PDF, for post-enumeration and audit.

Seven years of runs

Each run kept, and asked for one collection year at a time.

The statutory returns a school submits, each listing the parts of the record it reads, such as the census reading affiliations, census demographics, boarders and placements, and when it was last run. — Each return, and what it reads from the record. Invented data

Questions schools ask

Which returns does Kestrel produce?

The census with NCCD in the same file, the Address Collection, and STATS attendance, as SchoolsHUB files.

How does Kestrel help with post-enumeration?

Each run can produce a substantiation pack: the run record, the file’s checksum and, for each person, the rows behind their figures, as a PDF.

What happens when data is missing?

The run stops and names the student and the field, so the gap is fixed before the file is produced.

Show the detailHide the detail

A return is a run against a moment.

Each return declares which parts of the record it reads. Kestrel checks the record as at a moment, records what is missing by student and by field, and only then produces the file, with its checksum recorded on the run.

CensusStudents and full-time equivalent by campus, year level, sex, Indigenous status and boarding status, written as SchoolsHUB XML. A past run regenerates under its own year’s format.

NCCD, in the same fileThe adjustment levels and disability categories come from wellbeing’s own register, read over the same snapshot. The counts are in the census file and its checksum. They do not arrive as a second submission that has to agree with the first.

Address CollectionOne record per student, with their legal name and residential address, and a place for each parent holding parental responsibility: each with their own name and their own household’s address.

Checked weeklyEach open return is checked against the record each week, so the work of fixing it is spread across the term.

Census and returns, from the record. 0:38 · captions on screen

It stops and names what is missing.

Where the record cannot answer, the run stops and names the student and the field.

What stops an attendance returnTwo placements on one day. No campus. A sex or Indigenous status the school has not recorded. A value the collection year’s format has no code for. A mark recorded outside the student’s enrolment.

What stops an NCCD countAn adjustment with a gap in its evidence, one with no reason recorded, or a second row open at the same time.

A boarder’s term addressIs in neither field of the Address Collection, because it is neither the student’s residential address nor a parent’s household.

A protected recordThe address is withheld, the omission is named, and the file is still produced. A protected family should not be the reason a school misses a deadline, and should not be silently included either.

The substantiation pack.

This is the part that matters eighteen months later. Each run can produce a pack containing the run record, the file’s checksum and, for each person the run reported on, the rows behind their cells, written as a PDF an auditor can be handed.

The rows are read again, through the same parts of the record, as at that run’s own snapshot moment. Nothing was copied when the run was made, so a correction dated after the snapshot is correctly absent from the pack. The evidence is reconstructed as the record stood when the return was made.

That distinction is the whole difference between substantiation and a report. A school that exports today’s data to explain last year’s figure is producing a different number, and hoping no-one checks. An adjustment’s evidence pointers resolve to the student the adjustment was for, so an NCCD audit can follow the trail.

Producing a pack requires the passkey again, and the action is audited with the count of people it covers. Assembling each reportable fact about several hundred children into one PDF is exactly the action that should leave a mark.

Seven years of runs.

A return checked each week leaves hundreds of runs behind it over the years a school has to keep them, so the history is asked for one collection year at a time.

Producing a file also requires the passkey, and the file’s checksum is recorded on the run. So the question “is this the file we actually submitted?” has an answer that does not depend on somebody’s filing.

Where the STATS attendance rate comes from · The NESA and VASS submissions

Bring last year’s census.

Start with the eight-minute film. Then, if you’d like, forty minutes with us on your hardest figure, and where it came from.

Watch the walkthrough