This week, the results of the ninth round of the OECD’s Programme for International Student Assessment (PISA) were released.

PISA enables countries to track their educational performance over time. Since testing began in 2000, PISA has measured fifteen-year-olds’ performance in science, mathematics and reading. New Zealand’s results have declined substantially in all three subjects over those 25 years.

The 2025 data are not earth-shattering. No one PISA round ever is. It is the long-term trends that matter. Still, New Zealand’s results bring both hopeful news and reason for ongoing concern.

The best news is that the gap between students from advantaged and disadvantaged socioeconomic backgrounds has reduced markedly since 2022. This also happened in some other countries, but only because the performance of their most advantaged students fell. In New Zealand, the gap narrowed because the most disadvantaged students improved.

New Zealand showed a five-point improvement in science since 2022. For context, the OECD treats a twenty-point gain as representing a year of educational progress.

So, the science improvement represents about a quarter of a year’s learning. It is the first increase New Zealand has shown on any PISA scale since 2009.

Mathematics improved by just one point, too small a change to count as meaningful. Even so, the result is encouraging. New Zealand’s decline has been steeper in mathematics than in science or reading and, until now, New Zealand’s mathematics achievement has declined in every PISA round.

The total decline has been 43 points – equivalent to two full years of schooling by the OECD’s metric. While little can be made of the single point increase, New Zealand may at least have bottomed out. Moreover, New Zealand’s steady result cut against a nine-point drop in the international average.

Offsetting the improvement in science and stabilisation in mathematics was a decline of four points in reading.

This result may seem perplexing. Education Minister Erica Stanford mandated primary schools to teach reading using internationally proven methods from 2025. But education policy changes usually take years to show up in test results.  

It takes time to bed in changes to teaching practice, especially in a highly devolved school system like New Zealand’s. Students starting school under these reforms will not participate in PISA until the 2037 round.

PISA 2025 shows some green shoots for New Zealand education. But for flowers to bloom, successive governments must stay the course with Stanford’s reforms.

The four months before an election aren’t great for those who care about good policy.

Politicians’ competence at other times is questionable. During election season, it gets worse. Ridiculous policy proposals get airtime, with politicians hoping voters like the vibe. Just look at the past couple of weeks.

ACT promised to abolish my visa category. After causing a fair bit of panic, they clarified that the change would not apply to existing permanent residents. They may not have thought things through before the announcement.

New Zealand First said they’ll ban permanent residents from receiving NZ Superannuation. Most Western countries, including New Zealand, base eligibility on how long someone has contributed to the system. The policy would also break New Zealand’s reciprocal agreements with other countries.

Oliver’s column this week noted the Greens’ policy of nationalising supermarkets. At best, it will be a costly way for government to learn that running a supermarket chain is harder than it looks. But governments aren’t always the most diligent students.

National reaffirmed its commitment to social media age restrictions that would erode online privacy and freedom of speech. Every announcement that the government treats ‘big tech’ as an enemy makes it harder to convince international tech companies that New Zealand is a decent place to do business.

Perhaps it’s best that Labour has stayed fairly quiet.

Whatever your views on shifting to a four-year electoral cycle, it would at least reduce the share of each term consumed by election-induced insanity: 44 months out of 48 would be election-free, rather than the current 32 months out of 36.

Best not to follow elections too closely. Look instead to what people outside politics are achieving. We live in an age of wonders.

Mathematicians and AI are rapidly working their way through open, important, unsolved problems – many of which have stood for decades. New systems can also rigorously check that work. I can’t wait to see those techniques applied to difficult parts of engineering theory.

This week, Google released a free atlas predicting the effects of nine billion possible single-letter changes in the human genome. Meanwhile, personalised CAR T-cell therapies are curing cancers that would have been a death sentence only a few years ago.

Election campaigns are ugly things, but there’s no compulsion to stare into that abyss. The world is a beautiful place.

Look away, if you can.