When Phonics Isn’t Reading: Why Today’s ‘Literacy Breakthrough’ Doesn’t Add Up

This afternoon, the Minister of Education announced what she called a “transformational boost” in reading achievement, crediting the government’s mandated structured literacy programme and newly introduced phonics checks.

According to her statement, the proportion of five-year-olds “at or above expectation” on the 20-week phonics check rose from 36% in Term 1 to 58% in Term 3, while the number “needing support” dropped from 52% to 33%. She declared this as proof that structured literacy has “reversed decades of decline.”

It sounds impressive. But when you look at the data behind the headlines, a different picture emerges.

1. Phonics ≠ Reading

The phonics check is a short, 40-word decoding test designed to assess how well children can sound out words. It’s a diagnostic snapshot of one aspect of early reading: phoneme– grapheme correspondence. It is not an assessment of a child’s ability to read for meaning.

In other words, it measures whether a child can sound out a word, not whether they can understand it.

As any teacher knows, decoding is one thread in the complex weave of learning to read. True reading proficiency depends on comprehension, vocabulary, fluency, motivation, and a connection to text meaning. Without those, we have phonics performance, not reading success.

2. Two Data Points Don’t Make a Trend

The headline gains come from comparing different groups of children: those tested at 20 weeks in Term 1, Term 2, and Term 3, not the same children progressing over time.

The Ministry’s own report makes this clear:

“This data provides a cross-sectional view; progress of cohorts between timepoints is not yet available.”

That means these results show what this term’s five-year-olds could decode after 20 weeks, not how last term’s five-year-olds improved.

Statistically, this is called a cross-sectional sample, not a longitudinal cohort. It tells us nothing about growth, learning impact, or the effectiveness of any specific teaching programme.

To claim otherwise is misleading.

3. Seasonal and Sampling Effects Matter

Children tested in Term 1 have just returned from a two-month summer break; many have been at school for only a few weeks. By Term 3, most are well into school routines and benefit from continuity.

So, we’d expect later-term cohorts to score higher, even if teaching quality stayed the same.

The Ministry’s report also shows that the number of schools submitting results increased dramatically, from 194 schools in Term 1 to 458 schools in Term 3, and the total sample grew from around 1,100 to 4,300 children.

That change alone can shift percentages. Some schools joined because they were told phonics data was required to access Ministry support, an incentive that skews participation and results.

Put simply: when the sample quadruples and changes who’s included, the numbers cannot be used to show national progress.

4. Missing Data, Unclear Rules

The Ministry admits that some Term 3 schools used Term 1 question sets but were still included if two of three data points (test date, question set, submission date) matched the Term 3 criteria. That is not rigorous test equating.

Nor does the report show how many children were excluded or absent, or which schools declined to submit results.

This lack of transparency would not pass peer review in educational research. Yet these incomplete, shifting samples are being presented as evidence of “transformation.”

5. The Smallest Slice of True Growth

The only group tracked over time (children who completed both 20- and 40-week checks) was tiny: just 516 students or about 4.5% of the total sample. If we consider the broader total number of Year 1 students in 2025 is 68,349 – this figure represents 0.007 (or 7 in every thousand) of the current Year 1 cohort.

Even here, the Ministry’s report notes that results were “stable overall” and that while some students moved up a level, many remained within the same achievement band.

That’s not a revolution. It’s an expected incremental development in phonics knowledge.

6. Gaps in Equity and Comprehension

The Minister also claimed, “all groups benefited,” citing Māori students rising from 25% to 43% “at or above expectation,” and similar increases for Pacific students.

But the report itself acknowledges sampling skew: overrepresentation of European students and underrepresentation of schools with high Equity Index scores.

Without clear participation rates or confidence intervals, it’s impossible to confirm whether those apparent gains are real or statistical noise.

And critically, there’s no evidence yet of improved comprehension, writing, or vocabulary outcomes, the actual markers of literacy success.

We Need to Stop Equating Phonics with Literacy

Of course, no one disputes that instruction in grapheme-phoneme knowledge supports early decoding. But when governments reduce literacy to a single test of letter–sound knowledge (recall), they shrink the complexity of reading into a political headline.

The irony is that the same teachers now being praised for implementing “structured literacy” were already teaching phonic knowledge - often alongside rich oral language, play, and comprehension strategies that this new mandate risks crowding out.

Real Progress Requires Depth, Not Spin

If the government wants to claim success, it needs to release:

• Transparent participation data: which schools, how many children, and who’s missing.

• Matched cohort results tracking the same students over time.

• Complementary outcomes in comprehension and writing.

• Independent analysis reviewed by literacy researchers, not political advisors.

Until then, this is progress on a narrow phonics screen, not evidence of a literacy turnaround.

Evidence Before Celebration

Early literacy research does show that the first year of school matters and that strong foundations in phonemic awareness and phonic knowledge are an essential part of that picture. No one disputes that explicit teaching of sound–symbol relationships support early decoding.

But what the Minister’s announcement reveals is less about what works in classrooms and more about how evidence is being used to tell a political story. The data released this week does not yet show reading transformation at a national level; it shows a limited measure of one skill, drawn from a small and shifting sample, interpreted with far more confidence than the evidence can support.

Quality literacy instruction requires careful implementation, robust evaluation, and time to take root. When political storytelling outpaces sound analysis, it erodes confidence in the evidence teachers are told to follow and faithfully implement.

The gains in phonics data may well signal progress, but until we can link those improvements to broader reading outcomes, comprehension, and sustained literacy growth over time; and until improvements are demonstrated through methodologically-robust, independent, peer-reviewed research – it’s too soon to declare success.

Evidence should inform the story of education, not be shaped to fit it.