What state measures of student growth miss about real progress

Measuring how much students improve, rather than simply where they stand, has become the preferred way for state education departments to describe school performance. In New York, the state definition of student growth compares a student's change on the state assessment with the change of similar students elsewhere and reports the difference.

The logic is sound in outline: it should reward schools that move struggling children forward. The trouble is that a score built for system-level comparison does not describe any child, flattening the pupil who arrived two years behind and is now reading at grade level.

Australian families will recognise the pressure even where the names differ. National testing arrives through NAPLAN, university entry through the ATAR, and each state runs its own senior certificate — the HSC in New South Wales, the VCE in Victoria, the QCAA subject in Queensland. A number claiming to summarise a year of learning is hard to resist when it feeds league tables.

How a growth number gets built

State growth formulas run in three steps: estimate what a student should have scored given prior results, compare that estimate with what the student actually achieved, then average the surplus or shortfall across a cohort and publish it as a rating that administrators, parents and newspapers can repeat.

Each step carries an assumption. Cohorts need a minimum number of students before results appear, so small schools and small classes often go unrated. Students who move into or out of a district complicate the arithmetic, and their gains get credited to schools they no longer attend.

Once a definition becomes a funding formula and a public reputation, politics enters. New Yorkers United for Kids grew out of parents who wanted the state model examined rather than accepted, and similar instincts drive parent-led campaigns elsewhere in Europe, including the education positions advanced by the Free Voters' Party. Government sets the measure; the community should get to argue with it.

Where a cohort average loses the pupil

Schools usually read their own results by setting the accountability rating beside the raw test scores, and the two often point in different directions. Guidance on interpreting school quality data shows why participation rates, the movement of families in and out of the district, and the number of students who sat the day weigh as much as the scores themselves.

Opt-out campaigns sharpen the problem. When a meaningful share of parents decline the test, the remaining cohort stops being representative, and a growth figure built from it can flatter a school because of who stayed.

The child who improved most is rarely the child the report describes. A pupil who quietly caught up in reading may register no measurable growth because the test ceiling was never the barrier, while a newly diagnosed student may show a decline that reflects a change in support rather than lost capacity. Both look like failures on a spreadsheet.

Australia reads the same signals differently

Australia has no Common Core, but it has an equivalent: a national curriculum every state has agreed to, and a testing culture that turns results into school reputations. Measurement happens in state systems, so Victorian and Queensland students are counted with different instruments and then placed in the same national table.

The effects appear in familiar scenes. In Melbourne, try-out exams for selective schools are treated as a private market, and arguments break out over which schools deserve high-achieving children. In Sydney, catchments have become a property story, with buyers paying premiums for a postcode and parents measuring distance in train minutes.

The social glue is just as local: the sausage sizzle that raises money and keeps parents on the grounds after hours, the class heading to Byron Bay or the Gold Coast for Schoolies, and the parent committee that circulates this year's results against the neighbour's.

Reading the numbers without being defined by them

None of this makes test results useless. It makes them the answer to one question, and not the one families usually ask. In New York, parents who want the state model examined have gathered current action items, including guidance for county resolutions and for reaching a state senator. Every Australian state publishes comparable assessment and participation data.

A few habits make that scrutiny reliable:

What the number cannot say

The attraction of a standardised measure is that it can stand in for the thing itself, because a number is easier to argue with than a classroom. A growth score can show that a cohort, on balance, moved forward. It cannot show that a child is understood, that a family is supported, or that a particular skill has finally clicked.

That gap is where local knowledge lives, and it is why campaigns for local control of schools and parent opt-out networks keep pressing the same demand: governments should describe plainly what they measure, and leave the judgement of a child's progress to the people who teach that child. The fairest reading of a growth figure stays narrow. It describes a cohort on a given day, and says very little about the child in front of you.

✉