Does NAPLAN Measure Students, or Does It Also Change How Schools Teach?

Geoffrey Chen

When Numbers Start Making Decisions · Season One, “People on Either Side of a Threshold” · Article 4

1. What happens after a report enters the classroom?

Imagine that a primary school receives its latest NAPLAN results. The principal finds that the school’s Year 3 students performed below schools serving students with statistically similar backgrounds in writing. Several individual reports place students in the “Developing” or “Needs additional support” proficiency levels.

The school can respond in different ways. It can place the results alongside classroom work, teacher observation and students’ language backgrounds, then identify the writing capabilities that genuinely need strengthening. Or it can define next year’s objective simply as “raise the NAPLAN writing score”, increase template exercises, timed practice and repeated exposure to test formats, and temporarily displace parts of the curriculum that do not appear in the assessment.

Both responses begin with the same number. They produce different forms of teaching.

This shows that NAPLAN does not merely stand outside the classroom and record learning after it has occurred. Once results are used to set targets, compare schools, allocate support or evaluate work, they enter the causal chain of education. When the people being measured understand how they are measured, they may adjust their behaviour to fit the indicator.

The real question is therefore not whether NAPLAN can measure literacy and numeracy. It plainly provides relevant evidence. The question is what happens when that evidence acquires enough institutional weight. Is a school improving the capabilities its students genuinely need, or improving the performance that the test is able to see?

2. What does NAPLAN actually measure?

NAPLAN is Australia’s national literacy and numeracy assessment for students in Years 3, 5, 7 and 9. It tests reading, writing, conventions of language and numeracy. The Australian Curriculum, Assessment and Reporting Authority describes the results as additional information that supports teachers’ professional judgement, not a substitute for the extensive assessment that occurs in classrooms. ACARA also states explicitly that NAPLAN results do not measure the overall quality of a school. National Assessment Program: “NAPLAN”

The online reading, numeracy and conventions of language assessments use a tailored test design. Students begin from a common starting pathway, and the difficulty of later questions changes according to their earlier performance. The final result depends both on how many questions the student answers correctly and on the average difficulty of the questions assigned. This enables the assessment to cover a wider range of ability within a limited testing period, but it also means that an individual result is not simply a count of correct answers. National Assessment Program: “Tailored tests”

Since 2023, individual achievement has been reported using four proficiency levels:

  • Exceeding;
  • Strong;
  • Developing; and
  • Needs additional support.

“Strong” means that the student meets challenging but reasonable expectations at the time of testing. “Needs additional support” means that the student is not yet achieving the expected learning outcomes and is likely to need additional support to progress satisfactorily. The official explanation also cautions that these descriptions are fair summaries of skills generally demonstrated by students at a level, not complete descriptions of every individual. Students near the top and bottom of the same level may demonstrate noticeably different ranges and complexities of skill. National Assessment Program: “Proficiency level descriptions”

That qualification matters. A proficiency level is not a permanent identity. It is a classification generated by a limited assessment in particular domains at a particular time.

3. Where does the line between “Developing” and “Strong” come from?

The four levels can look like naturally occurring stages of learning. In reality, they require cut-points chosen through a standard-setting process.

The NAPLAN 2025 Technical Report explains that the new thresholds introduced in 2023 were established through three linked steps. Panels of experienced teachers, curriculum experts and assessment specialists conducted standard-setting; statistical analysis tested whether the cut-points produced a smooth growth trajectory from Year 3 to Year 9; and the skill descriptions associated with each proficiency level were validated. ACARA: “NAPLAN 2025 Technical Report”

The thresholds were therefore not drawn arbitrarily. They combine professional judgement, curriculum expectations and statistical analysis to answer a practical question: which students have reached a challenging but reasonable expectation for this point in schooling, which students are still developing, and which students are likely to require additional support?

Yet “set through a rigorous process” does not mean that a fundamental difference exists on opposite sides of a cut-point. The official proficiency descriptions themselves acknowledge variation within each level. A student who falls just inside Developing and one who falls just inside Strong may possess very similar underlying capabilities even though the institutional label changes.

The most defensible use of this boundary is to trigger further observation and appropriate support. Its most dangerous use is to turn one classification into a final judgement about a student’s ability, a teacher’s quality or a school’s value.

4. When an indicator becomes a target, what happens to the thing being measured?

NAPLAN results have several legitimate uses. Parents can use them to discuss a child’s strengths and needs with teachers. Teachers can identify students who may need greater challenge or additional support. Schools can investigate gaps in their programs. Governments and education systems can monitor persistent differences across groups and regions.

But when the same indicator becomes connected to reputation, resources and accountability, schools acquire an incentive to optimise it.

This feedback does not require dishonesty. If teachers see that students are consistently weak in a particular kind of writing task, adjusting instruction may be entirely appropriate. The difficulty is that an indicator can be improved along two different paths:

  1. improve the underlying capability represented by the indicator, allowing the score to rise with it; or
  2. train students to respond more efficiently to the specific assessment, raising scores while producing a narrower improvement in capability.

The short-term data may not reveal which path has been taken.

An earlier Australian Government review of ACARA recorded concerns about excessive reliance on testing, student stress, “teaching to the test”, narrowing the curriculum, privileging literacy and numeracy over other learning, and placing minimum competencies above creative and analytical capabilities. Those concerns do not prove that every school responds this way. They identify a more general risk: the greater the consequences attached to a public indicator, the more organisational attention flows towards the parts of education that the indicator can capture. Australian Government: “Review of the Australian Curriculum, Assessment and Reporting Authority”

A number originally created to observe the education system can therefore begin to reshape that system. The assessment no longer asks only, “What have students learned?” Schools also begin to ask, “Which kinds of learning will appear in the assessment?”

5. A school’s results are not the same as the school’s contribution

Differences in school-level NAPLAN results create another easily overlooked problem. Student achievement is influenced by the school, but also by family resources, language background, prior learning, health, regional conditions, student mobility and many other factors.

My School provides school results and tools for comparison with schools serving students from statistically similar backgrounds. This is more meaningful than directly comparing two schools whose communities are profoundly different. But “statistically similar” is not a randomised experiment. It cannot remove every unobserved difference. A higher average result does not show that the school caused the whole difference, and a lower average result does not prove that teachers failed to produce progress. ACARA: “My School”

Averages can also conceal distributions. The same school average could arise because most students are clustered near the middle, or because one group performs very strongly while another requires intensive support. If a resource decision examines only the single average, the students most in need may once again disappear inside the number.

NAPLAN can therefore serve as a signal for school improvement, but it cannot carry an entire school rating by itself. ACARA’s statement that NAPLAN does not measure overall school quality is not a polite disclaimer at the edge of the system. It is a central boundary on what the indicator can validly support.

6. If a result appears inaccurate, who can review it?

Different kinds of problems can arise during NAPLAN. A student may be ill on the day of testing. A device or network may fail. Required disability adjustments may not be provided correctly. A test administration incident may occur. A family may also believe that the result is sharply inconsistent with the student’s long-term classroom work.

Technical and administrative incidents are addressed through the school and the relevant state or territory test administration authority. When parents have not received a report or have questions about a result, official guidance directs them first to the school. For the educational judgement itself, the most important review is not a demand to move a borderline score into the next proficiency level. It is to compare the NAPLAN result with classroom work, school-based assessment, teacher observation and, when necessary, specialist assessment. National Assessment Program: “NAPLAN results, reports and performance”

These two forms of review should not be confused:

  • review of a testing incident asks whether the data and procedure were correct; and
  • review of an educational judgement asks whether one result is sufficient to support a decision about a student or about teaching.

If NAPLAN remains one additional source of evidence, an unusual result can be corrected or qualified by other information. If it is treated as the only truth, a decision can be wrong even when the score has been calculated perfectly.

Conclusion: good measurement should improve teaching, not replace its purpose

NAPLAN gives Australia’s education system a common language of observation. Without shared measurement, long-running disparities can be concealed by local impressions, and students or schools needing support can be harder for systems to identify.

But NAPLAN measures only the part of capability that can be demonstrated through standardised tasks in selected domains at a particular time. It does not measure all of a school’s educational value. Nor does it exhaust a student’s knowledge, creativity, curiosity or conditions for growth.

The real judgement is therefore not whether testing should exist. It is how much weight test results should carry in subsequent decisions:

NAPLAN is suited to being the beginning of further observation and support. It is not suited to being the end point of a total judgement about a student, teacher or school.

An educational indicator is most successful when it does not make every classroom increasingly resemble the test. It should help teachers identify problems that were previously difficult to see, after which a rich curriculum and professional judgement take control of teaching again.

When a number begins changing the people and institutions it measures, the system must continue measuring a second outcome: have we improved students’ capabilities, or merely improved the indicator?

Primary sources

Discover more from Geoffrey Chen

Subscribe to get the latest posts sent to your email.