At first glance, the territory tops the nation in most of the subjects and age groups tested each May in the national assessment. Yet as soon as you compare children living in metropolitan areas or those with university-educated parents - a more accurate reflection of Canberra's population - the ACT plunges down the rankings.

In September, leading think tank the Grattan Institute delivered the latest in a series of sobering report cards for the ACT, finding the territory was trailing the nation in student learning progress despite its socioeconomic advantages. It follows recent analysis by the Australian National University which revealed a "sea of red" where Canberra high school students were lagging up to 16 months behind their peers from similar backgrounds, and earlier work from the ACT auditor-general that showed students were underperforming despite government spending.

The ACT government, which has attempted to discredit findings from both the Grattan Institute and the ANU in recent weeks, insists any school performance problem in the territory is one of equity.

But experts argue it goes far beyond socioeconomic divides in a system where the vast majority of schools enjoy above-average levels of advantage.

In August, Education Minister Yvette Berry unveiled her much anticipated 10-year plan for education - a strategy promising to boost performance by lifting students onto a level playing field. The recent roll-out of Chromebooks in ACT schools was billed as an early example of this approach, placing technology in the hands of every student regardless of their family's means.

"Even in wealthy communities like the ACT children start life in vastly different places," Ms Berry said.

A 2017 report found 12 per cent of Canberra children were living in low income households, while five per cent of the population did not complete Year 10 and were no longer in school (compared to an average of 10 per cent for other capital cities).

Alongside community consultation, the government said the driving piece of evidence behind its new equity push was a robust analysis of 2013 NAPLAN data by leading education expert Stephen Lamb.

Professor Lamb found Canberra children "on average achieve negative results on every measure" compared to those of similar socioeconomic status, but his report had also outlined the impact of an apparent class divide in the sector, as aspirational families moved their high-achieving children to popular schools outside their local area.

"[He found] the ACT performance issue is one of equity," Deb Efthymiades from the Education Directorate told a recent assembly inquiry.

In an affluent place like Canberra, where disadvantage is not absent but instead spread unusually thin, Professor Lamb said it was interesting to see equity still having an impact on performance.

But he also said his report appeared to have been misinterpreted by the directorate, telling The Canberra Times an equity problem was no more to blame for poor academic results in Canberra than in any other school system across Australia. He had included it in the report to help sketch out the state of the school system but said his remit from the ACT government had been to assess the sector's performance, not find the reasons behind poor results.

Ms Berry has since clarified that, while the report had been important to understanding 'what's happening" in ACT schools, the strategy was also informed by research from around the world. Internationally, the highest-performing school systems combined quality with equity, she said.

ACT head of the Australian Education Union Glenn Fowler says, given Canberra has the most highly privatised education system in Australia, it's no surprise schools are feeling the effects of residualisation as parents shop around.

But while researchers agree closing the gap between schools is crucial to boosting any system's performance, Peter Goss at the Grattan Institute says equity appears to be less of a problem in the ACT than in other jurisdictions.

His most recent analysis found more cause for concern among the capital's advantaged students, which lag behind other states, than in its relatively disadvantaged cohort. The same trends in underperformance also appears in the non-government sector, though to a slightly lesser degree.

At the University of Sydney, education assessment expert Rachel Wilson says that while residualisation and segregation between schools is a problem across Australia, as pointed out by OECD reports, it does not explain sliding standards across the board.

"Besides equity is not as big a problem in the ACT," she says. "You need to look closer than that."

Literacy expert Misty Adoniou offers her own theory. She says it's possible ACT classrooms have a wider range of needs than in other states, where children from high or low socioeconomic backgrounds tend to be concentrated in particular schools, making it easier to tailor teaching to them.

"But that's not an excuse," she says of the ACT system.

"l'm always concerned every year when the ACT trumpets that it's topped the NAPLAN...what's actually happening? Why it is that we're not as good as we should be?"

Associate professor Simon Leonard points to the ACT's large gap in results between low and high SES students. As disadvantaged pupils are scattered between schools in smaller numbers than elsewhere in Australia, a school's relatively high average performance often hides the problem, he says, allowing complacency to set in.

Trevor Cobbold, former Productivity Commission economist and convenor of the national lobby group Save our Schools, agrees the ACT could do more for disadvantaged students.

While the territory has a good record retaining students until Year 12, in the most recent productivity commission report it also had the nation's third-lowest completion rate among low SES students. Unadjusted for socioeconomic status, international PISA scores place the ACT at the top of the pack for academic performance in Australia - though falling flat when it comes to disadvantaged students and student sense of belonging.

But Mr Cobbald remains sceptical of the ACT's equity strategy, which he says has yet to present any practical benchmarks or detail.

At the other end of the spectrum, Dr Goss warns the ACT is not good at stretching its top students, particularly in numeracy.

"It's good in reading and it's actually better at supporting students from disadvantage relatively across board," he says.

Dr Adoniou suggests that trend could be due to the fact more ACT students are already at the higher end of results, where it can be trickier to make progress compared to at lower levels.

But Dr Goss says it's not a case of students performing well and then slowing down, children at all levels make less progress in the ACT than elsewhere around the country.

Dr Andoniou points to another sting in the NAPLAN data. By Year 9, the number of ACT students achieving minimum standard drops away considerably compared to Year 3.

Schools across Australia are focussed on getting students over early benchmarks, she says, but at its higher levels NAPLAN tests for creativity and originality of thought more than following a formula, particularly in writing.

"We’ve really forgotten that top end...and we've forgotten to stretch the low-achievers too," she says.

"It's time to get back to the magic and the power of writing."

A 2017 ACT auditor-general report also questioned the territory's practice of giving schools lots of autonomy - something it called "excessive" for a small jurisdiction.

The ACT's peak parents' group says resulting inconsistencies between schools has created the appearance of a class divide in the sector.

While Dean of Education at the University of Canberra Geoff Riordon agrees ACT teachers have far more freedom in the classroom than elsewhere around the country, he says that's one of the territory's strengths as a school system.

At Grattan, Dr Goss says school autonomy on its own appears to have less of an impact on performance than teaching methods.

Some researchers have called for targeted interventions to be rolled out at a classroom level, but the union is wary of any shift to more standardised methods, saying teachers are best placed to determine what their students require.

Still, Mr Fowler is adamant about the need to lift the standard of teacher trainees. The union wants a minimum ATAR of 70 to study teaching at university, and has pushed for longer courses.

Across the country, Dr Wilson warns some universities are accepting students with ATARs in the low 20s (Mr Fowler reports 31 here in the ACT). While international benchmarks say teachers should be graduates from the top 70 per cent of their year level, she says the majority enter a teaching degree with an ATAR below 50.

"A lot of our gaze in education has been downward on schools but the problem is not in schools, it's in the system, it's in the universities who control quality," she says.

"One in five initial teacher programs is being taught online. We're not attracting the graduates, we've failed to roll out a proper recruitment policy - although people like Eddie Woo [the maths teacher] are helping."

At the territory level, Grattan analysts also point to PISA surveys which report Canberra teachers are more resistant to change than in other parts of the country (with the exception of Tasmania) and clock up higher than average levels of absenteeism.

Last year, the auditor-general found Canberra teachers were not collecting and analysing school performance data effectively, which Grattan's Julie Sonneman says could make it hard for teachers to pitch lessons to the best advantage.

Then there are those who argue the problem could be NAPLAN itself, as it tests too narrow a skill set in literacy and numeracy to accurately measure a school system.

The controversial national assessment - and the publication of its results - has come under increasing fire from ministers, teachers and academics who claim it has created a "high stakes" culture of standardised testing in Australia.

Comparatively low participation rates in the ACT, particularly among Year 9 students, and the unique spread of the capital's disadvantage are sometimes held up as evidence of the unreliability of results in such a small jurisdiction where only about 4000 students sit the test each year.

Professor Leonard urges caution in over-interpreting the data, saying small differences, even down to the weather, can make all the difference in Canberra.

"I suspect that the higher incomes found [here] do not translate to high levels of social and cultural capital as they would in other places," he says.

"The public sector keeps middle incomes in Canberra high and perhaps you have to adjust for that.

"Even so you'd expect the maths scores in the ACT to be better."

Andrew Macintosh, a policy analyst and co-author of the ANU report, acknowledges the problem might come down to the data itself, but adds you would expect to see a more consistent pattern in that case. Instead, ACT students do much better in reading than in writing and numeracy.

Early work has begun at ACARA, the agency which runs NAPLAN, to review its socioeconomic calculations in relation to the ACT, after an approach by the ACT government.

Minister Berry said government analysis had found the even spread of disadvantage across ACT schools affected "both how it is presented in performance data and the strategies required to address it".

At the union, Mr Fowler says measures beyond NAPLAN such as ATAR scores and later employment should to be taken into account, but agrees the ACT needs to look "under the bonnet" of its school system, across both public and private sectors.

"I'm not convinced there's a problem [based off NAPLAN] but I don't think this should be buried either."

While the Grattan Institute concedes NAPLAN does not capture everything that matters in education, it stresses "it is the only test in Australia that enables us to compare student progress across every school".

Professor Macintosh also defends the test as the best and most cost effective tool to measure any intervention into the problem in Canberra's classrooms - and hold systems to account.

Mr Cobbold, who admits he has become more wary of NAPLAN in recent years, says the test would likely produce more reliable results across the board if it was more engaging for students.

Dr Andoniou agrees, adding: "Imagine if the NAPLAN [writing testing] was one day in May when fifteen-year-olds got the chance to share their ideas with Australia - and we actually listened!"

with Kirsten Lawson