Manipulation of NCEA data is giving some parents the wrong idea about so-called "miracle schools", the former face of NCEA says.

Ellie and Courtney McManaway were all smiles after successfully passing their 2014 NCEA exams. Photo: Tracy Hardy.

Bali Haque, who was deputy chief executive of the New Zealand Qualifications Authority, said there "could be some schools who are abusing the system".

The problem is with schools said to be inflating their "participation-based" pass rate without disclosing to parents less favourable "roll-based" results.

Nationally, the gap between the participation and roll-based pass rates is typically about 11 percentage points at NCEA levels one and two and about 19 points at level three.

But Stuff.co.nz has identified dozens of schools where the gap is much wider and getting bigger at every level. Some of those schools are reporting only their participation-based figures to parents and claiming to have made "dramatic gains".

In some cases, participation-based pass rates have climbed dramatically to almost 100 per cent but roll-based pass rates have languished at just over half, or less. The participation-based measure does not count students who have backed out of NCEA during the school year, whereas the roll-based method counts all of the students at a school.

"You could have some schools who are abusing the system, and they're not entering students because they think students are going to fail. That manipulates the data and is an abuse of the system," Haque said.

Edgecumbe College, a decile three Eastern Bay of Plenty school visited last month by Prime Minister John Key, has had a steadily increasing gap between its participation-based and roll-based pass rate since 2008.

On the participation-based measure, pass rates were between 83 and 100 per cent last year. On the roll-based measure, however, pass rates at the school were between 33 and 64 per cent.

"Our students sometimes take longer to get there than other students. So we're not putting them through so they fail one standard after another, after another," the deputy principal of Edgecumbe College, Anthony Murphy said.

Haque said parents needed to be clear about what was actually going on at schools, and not use NCEA results to "jump to conclusions about a 'miracle' school with an outstanding principal who has finally 'got the answer'".

"Sometimes you read in the papers about a particular school that's increased its NCEA pass rates by 30 or 40 per cent in a year or two. Immediately your mind goes to, what's going on there?" Haque said.

"I call them heroic schools – suddenly it just happens. I want to know why. And if you know why, then it becomes something we can understand, rather than something we should be suspicious about."

Education Minister Parata said she was "emphatically" focused on roll-based data.

"All of us communicate the best about ourselves and the best about our businesses and in this case, about our schools," she said.

"But as Minister of Education and through the Ministry of Education, and through ERO's reviews, it is about roll-based data because we want to understand what is happening with every child - not just the ones that are being successful.

"I've had this debate with principals and peak associations and I'm absolutely unmovable on the fact that the Government uses roll-based data."

Secondary Principals' Association president Sandy Pasley said people should be asking questions if schools had more than a five or six percentage point difference between their figures.

"Is it because you have a lot of students leaving school, or is it because you've got a special education unit associated with your school? Or are you just not entering them because you don't want to upset your figures?

"It's a bit of a minefield, working out what's happening."