A Canadian principal opens the latest international assessment report and sees two statements that appear to contradict one another. Canadian students still outperform the OECD average in science, mathematics and reading. Yet the country's performance has fallen to its lowest recorded level in all three subjects. Should the school celebrate or sound the alarm?
The useful answer is neither complacency nor panic. PISA 2025 shows a strong Canadian education system facing a persistent learning challenge. Schools should protect what works while examining how students read, reason, explain evidence, monitor understanding and seek help. That means treating international scores as signals for investigation—not verdicts on individual teachers or children—and translating the findings into better everyday learning.
The OECD's Canada country note and the Council of Ministers of Education, Canada's PISA 2025 materials provide the evidence. What follows distinguishes their reported results from our practical recommendations for Canadian schools.
What did PISA 2025 actually find in Canada?
Released on 8 September 2026, PISA 2025 assessed 15-year-old students in science, mathematics and reading. Science was the principal domain. The cycle also examined computational problem-solving and gathered students' and principals' accounts of learning environments. Approximately 24,600 Canadian students across 897 schools participated, according to the OECD.
Canada's central finding is a double reality: results remain above the OECD average in the three established subjects, but achievement in mathematics and reading declined compared with 2022. Science performance was roughly unchanged over that shorter interval. All three 2025 subject results were lower than Canada's earlier PISA results and significantly below 2015 levels.
Those are not contradictory statements. A country can retain a comparative advantage while its own learning outcomes deteriorate over time. For school leaders, the second comparison may be the more urgent one: Are today's students developing the competencies Canada's earlier cohorts demonstrated?
The proficiency numbers behind the headlines
Instead of beginning with national rankings, examine the share of students reaching PISA's baseline, Level 2. These figures provide a clearer starting point for discussing the breadth of achievement.
| PISA 2025 domain | Canada at Level 2 or above | OECD average at Level 2 or above | Canada below Level 2 |
|---|---|---|---|
| Science | 84% | 74% | 16% |
| Mathematics | 74% | 65% | 26% |
| Reading | 78% | 69% | 22% |
Source: OECD, PISA 2025 Canada country note. Figures are rounded; the rightmost column is the complement of Canada's Level 2-or-above share.
These figures are simultaneously encouraging and demanding. Many Canadian students demonstrate the basics assessed by PISA. But roughly one in four students did not reach baseline proficiency in mathematics, and more than one in five did not do so in reading. Such proportions deserve sustained attention even in a country that performs relatively well internationally.
Science also offers a useful warning against simplistic interpretations. Baseline proficiency does not mean full command of the discipline, and a Level 2 threshold is not a school graduation standard. It identifies capabilities demonstrated in a particular international assessment. Schools must still examine their curricula, local assessment evidence and students' actual work.
What has changed since 2015?
The OECD reports that, compared with 2015, the share of students below Level 2 increased by 12 percentage points in mathematics, 11 points in reading and five points in science. These are percentage-point changes, not percentage changes or claims about individual students.
The distance between the highest- and lowest-scoring Canadian students also widened in mathematics from 2022 to 2025. That is a reason to investigate who is progressing, who is not and which classroom supports are reaching learners—not a reason to assume one policy caused the change.
The lesson is uncomfortable but productive: maintaining a high international standing is not the same as maintaining the depth and distribution of learning over time.
Before changing school policy, understand what PISA cannot prove
PISA is a sample-based international assessment of 15-year-olds. It is not an annual test of every Canadian school, nor does it measure the impact of a particular classroom intervention. Questionnaire responses offer important context, but associations between attitudes, conditions and scores do not establish cause and effect.
The OECD also flags a significant data-quality limitation for Canada. Student and school response rates fell below PISA's technical targets. Its analysis indicates that student non-response produced a slight upward bias in national results; scenario analysis suggests that this upward bias in science is unlikely to exceed 15 points. A responsible analysis acknowledges this qualification rather than quietly converting a national estimate into false precision.
Canada's education systems also differ by province and territory, language, community and school context. A national average cannot specify which intervention is appropriate in a rural northern school, an urban high school, a French-language minority setting or a classroom welcoming newly arrived multilingual learners. The Canadian PISA report and provincial comparisons should therefore be read beside local evidence.
Three interpretations are especially risky:
"Canadian schools are failing." The evidence does not support such a sweeping claim; Canada remains above OECD averages.
"Canadian students are fine because the country ranks well." Relative success does not erase long-term decline or unmet learning needs.
"Technology, parents or teachers caused the decline." PISA describes patterns and relationships; it does not isolate a single cause.
Better questions are more specific: Which competencies are weakening? Which groups or contexts require additional support? What do students do when an unfamiliar task resists their first attempt? And what evidence would show that a school response is helping?
The overlooked signal: students need to learn how to learn
One of the most actionable parts of the 2025 country note is not a headline score. It concerns students' approaches to their own learning.
The OECD reports that approximately 45% of Canadian students say they often or very often plan how they will study. Fifty-six percent say they ask themselves questions to check their understanding. Meanwhile, 81% report seeking help from teachers when necessary, and 83% from classmates when they do not fully understand.
These responses describe reported habits, not verified mastery. Still, they raise a question every Canadian school can investigate locally: Do we explicitly teach students how to plan, check and adjust learning, or do we assume those skills will develop automatically?
A student may be given excellent resources but have little idea how to detect a misconception. Another may revise for hours by rereading notes, unable to decide what deserves practice. A third may complete an assignment with substantial adult prompting and then struggle on an independent transfer task.
The goal is not to label these students unmotivated. It is to identify the precise skill needed next.
From answer production to evidence of understanding
Consider a Grade 9 science lesson in which students must evaluate a claim about water quality using a table and two short source texts. The teacher could simply demonstrate the answer and ask the class to copy the reasoning. Everyone might leave with a completed worksheet.
A stronger sequence asks students first to identify the claim, the evidence, the uncertainty and any missing information. After an initial attempt, the teacher models how to distinguish a plausible statement from one supported by the data. Students then try a new example independently and explain why one source is more reliable than another.
The important measure is not whether the first worksheet looks complete. It is whether students can transfer the reasoning to unfamiliar evidence.
This is the shift from answer delivery to learning coaching: teach substantive content clearly, then use questioning, feedback and gradual release so learners become increasingly responsible for their own thinking.
Science literacy: teach students to examine claims, not repeat conclusions
Science was PISA 2025's principal domain, making it a natural starting point for school discussion. The Canadian report describes work involving explanations of phenomena, scientific inquiry, critical interpretation of data and evaluation of information for decisions.
Those demands fit a world in which students constantly encounter graphs, health claims, environmental debates and AI-generated explanations. Merely remembering the correct technical vocabulary is insufficient when two sources disagree or when a graph appears persuasive but omits a crucial scale.
A practical science routine can follow four moves: claim, evidence, alternative, decision. Students state what is being claimed, identify the observations that support it, test a competing explanation and explain what they would conclude provisionally. Teachers can adapt the demand to age and prior knowledge.
For instance, learners might compare two community proposals to reduce a school's energy use. One includes cost estimates and measurements; the other relies mainly on slogans. Rather than rewarding the recommendation the teacher prefers, assess how clearly students justify their decision with evidence and identify uncertainty.
Scientific literacy is not the exclusive responsibility of science departments. History students evaluate sources. Mathematics students question graphs. English language arts students distinguish fact, inference and persuasion. A school-wide emphasis on evidence can connect these disciplines without imposing identical activities on every class.
What should teachers look for?
A learner who understands a topic can often restate it accurately, explain the reasons behind a conclusion, apply the idea in a new situation and notice when a result is implausible. None of these behaviours must occur perfectly on the first attempt. Their development requires feedback and repeated opportunities to practise.
The school leader's question should be: Where, across the timetable, do we ask students to explain how they know? If that moment appears only in occasional examinations, the curriculum may offer too few opportunities for genuine reasoning.
Reading and mathematics: diagnose the problem before increasing practice
Declining reading and mathematics results understandably produce calls for more instructional time. But more of the same practice is not automatically the right response.
In reading, a student may decode words accurately while struggling to infer an author's purpose, connect ideas across paragraphs or evaluate sources. Another may understand a text when read aloud but be slowed by fluency or vocabulary. These needs are not interchangeable. An appropriate school response begins by inspecting tasks and student responses to distinguish the obstacle.
Teachers might ask students to annotate two brief texts that take different positions on a community question. Rather than asking only for the topic, invite learners to identify the strongest supporting sentence, explain an implied assumption and compare the credibility of the evidence. Targeted vocabulary and comprehension instruction should precede independent evaluation where necessary.
In mathematics, one learner may lack a foundational procedure; another may execute procedures successfully but fail to recognise when to use them. A familiar worked example can hide the second problem. The next step is a related, unfamiliar task requiring students to choose a method and justify why it fits.
For example, instead of ten nearly identical percentage exercises, ask students to compare two promotional offers and explain which is better under different assumptions. The class still needs explicit instruction and practice in calculation. It also needs opportunities to make mathematical decisions.
Use three kinds of evidence together
A school should combine curriculum assessments, student work samples and short conversations about reasoning. The first shows what was answered; the second reveals how; the third can expose misconceptions that a score conceals.
Before purchasing a new program or adding hours of homework, staff should ask whether a common misconception requires explicit reteaching, a prerequisite needs strengthening or students need more independent transfer opportunities. The instructional response will differ in each case.
Equity matters even when the average gap looks comparatively small
The OECD reports a 59-point gap in science between Canadian students in the highest and lowest national socioeconomic quarters, compared with an 85-point average gap across OECD countries. This comparatively smaller difference is not evidence that socioeconomic disadvantage has disappeared.
National aggregates can hide important differences by province, community, language of instruction, disability-related access needs and opportunities to learn. It is equally important not to treat any such characteristic as a fixed prediction of ability.
School teams can examine whether learners have equitable access to appropriately challenging texts, trained staff, laboratory experiences, technology, quiet study spaces and timely help. They can ask whether assessment formats allow students to demonstrate the intended skill without unrelated barriers.
In multilingual settings, language development and subject understanding sometimes need to be assessed separately. In remote communities, digital resources may offer opportunities while connectivity and staff capacity remain constraints. For Indigenous students, culturally responsive learning and partnerships with communities should inform local decisions rather than be reduced to a comparison of test scores.
Equity is not achieved by lowering expectations. It requires making the support, teaching and opportunities necessary to reach demanding goals genuinely available.
Digital devices and AI: focus on thinking rather than screen time alone
PISA 2025 adds timely evidence about school technology. In Canada, 38% of students reported using AI chatbots for learning at least weekly. The OECD also reports that 32% said peers were frequently distracted by digital devices during most or all science lessons. These figures describe student reports, not an experiment demonstrating the effects of AI or phones.
That distinction matters. Technology can support access, feedback and exploration, but it can also make it effortless to produce a convincing answer without understanding it. Banning all tools or embracing every new platform misses the instructional question: What thinking remains the student's responsibility?
Consider a carefully bounded AI exercise. A student asks a chatbot to explain a scientific claim, checks the explanation against the class's verified sources, marks any unsupported statements, and writes a revised answer without copying. The teacher evaluates the source checking, reasoning and revision—not merely the final polished paragraph.
The reverse activity is just as useful: solve the task independently first, then use AI as a critic. The student must decide whether to accept any suggestion and explain why. Schools should also set age-appropriate privacy safeguards and follow applicable provincial policies before introducing tools.
Digital literacy is therefore part of learning literacy. It includes knowing when assistance helps, when it masks a gap, and how to verify what appears authoritative.
A practical 90-day response for Canadian school leaders
The following is a proposed improvement framework, not an OECD-endorsed intervention or a proven Canadian pilot. It can be adapted by a school, district or board according to resources and local requirements.
Days 1–30: identify one high-value learning problem
Begin with an honest review of available evidence: provincial assessments, current classroom work, student voice, attendance patterns and support needs. Separate achievement concerns from issues of access or school climate. Select one problem specific enough to observe—for example, Grade 8 learners identifying evidence in science texts but struggling to justify conclusions.
Invite teachers who know the learners to test whether that description holds in ordinary classroom tasks. Ask students which part of the task feels difficult. Collect a small baseline of anonymised work samples using consistent criteria. Avoid using PISA alone to diagnose your school.
Days 31–60: make the skill teachable and visible
Co-design a short instructional routine with teachers. Model the target thinking explicitly, provide supported practice, then introduce an unfamiliar application. Agree on what counts as improvement before judging results: a stronger claim-evidence connection, a clearer explanation, a sounder choice of method or better independent planning.
Hold brief, supportive professional conversations centred on the task rather than a teacher's personality. Questions might include: Where did learners struggle? Which prompt helped? What could we remove next time so students do more themselves?
Days 61–90: test transfer, refine and decide
Revisit comparable tasks and student work with the same rubric. Check whether the gain appears only on rehearsed examples or on new problems too. Examine whether the change benefits students who previously struggled, not simply the class average.
Ask students whether they can explain the learning strategy they used and when they would choose it again. If evidence is weak, revise the routine or reconsider the initial diagnosis. If promising, continue monitoring before widening it to other departments.
| Question for the team | Practical evidence | Decision it informs |
|---|---|---|
| Can students explain a claim with relevant evidence? | Before-and-after annotated work | Continue or revise explicit reasoning instruction |
| Can they apply the method to a new problem? | Unseen transfer task | Reduce or retain scaffolding |
| Do they monitor their own understanding? | Brief learning reflection with an example | Teach a more concrete self-check routine |
| Who is benefiting and who is missing out? | Disaggregated, privacy-protected work review | Adapt access and support |
| Are staff able to sustain the change? | Teacher feedback and workload check | Simplify or resource the approach |
Such a cycle turns a distant international assessment into a local inquiry without pretending that 90 days will alter national scores.
The teacher's role: clear instruction and coaching are partners
There is a false choice in education debates: either teachers explain knowledge or students discover everything by themselves. Effective learning needs both guidance and opportunities for independence.
A learning-coaching approach does not mean withholding answers from a child who lacks the prerequisite knowledge. It means making the purpose of support explicit and deciding when to move responsibility from teacher to learner. Early in a lesson, modelling may be essential. Later, carefully framed questions may help students test their understanding.
Consider three teacher prompts:
Before: “What is this task asking, and what do you already know that might help?”
During: “Which part is certain, which is an assumption, and how could you check?”
After: “What would you try first on a different problem, without my reminder?”
These questions work best when grounded in accurate subject knowledge. They are not a replacement for explicit reading instruction, mathematical explanation, specialist assessment or appropriately structured support.
For educators exploring this professional posture, Learning Coach describes pathways for teachers and educational leaders on its training and specialization page. The idea is to connect observation, questioning, practical strategies and learner autonomy. Schools should review the current details and availability of a specialization before applying; an application does not guarantee admission or certification. Educators interested in a fuller discussion of the teacher's evolving role can also explore Stop Teaching, Start Coaching.
What can parents do without turning home into another classroom?
Parents do not need to imitate PISA test questions or supervise every minute of homework. Small shifts in conversation can make learning more visible.
When a teenager returns with a completed assignment, try asking, “Which question made you think hardest?” or “How did you know your answer was reasonable?” If they are stuck, ask what they have already tried and what specific help they need. Offer a worked example, a clearer explanation or a brief break when appropriate; independence is not abandonment.
For reading, discuss an article's central claim and what makes a source trustworthy. For mathematics, invite the teenager to explain why a method applies before checking the arithmetic. For science, practise separating observation from interpretation in ordinary news and community questions.
Avoid interpreting a single disappointing test result as a character flaw. Families can work with teachers to identify foundational gaps, access accommodations or additional instruction when necessary, and celebrate the effort of trying a better strategy—not merely finishing quickly.
Five decisions Canada should resist—and five it can make instead
An international assessment generates pressure for visible action. But a visible response is not necessarily an effective one. A school or province should resist the temptation to substitute slogans for careful diagnosis.
Resist ranking obsession. Instead, ask whether more learners are demonstrating meaningful proficiency.
Resist blaming one group. Instead, bring educators, learners, families and community voices into a focused problem-solving process.
Resist technology theatre. Instead, require every digital tool to demonstrate its contribution to authentic thinking, access or feedback.
Resist endless drilling. Instead, combine secure fundamentals with unfamiliar applications where students must choose and justify an approach.
Resist asking teachers to change everything at once. Instead, give teams time, evidence and practical professional support to improve one consequential learning routine.
Canada's provincial and territorial responsibilities for education make a single national prescription inappropriate. Shared evidence and professional collaboration, however, can make local action more intelligent.
Frequently asked questions about Canada's PISA 2025 results
Did Canadian students perform below the OECD average?
No. In PISA 2025, Canada performed above the OECD average in science, mathematics and reading, even as its own results reached historic lows relative to previous Canadian PISA cycles. Both comparisons are true, and both matter.
Did science achievement fall between 2022 and 2025?
The OECD describes Canada's average science results as about the same as in 2022. Mathematics and reading were lower. Longer-term comparisons to 2015 nevertheless show significant declines in all three subjects.
Is PISA a measure of every school in Canada?
No. It is a sample-based assessment of 15-year-olds. National or provincial findings cannot diagnose an individual school or student without additional evidence.
Does PISA prove that phones or AI caused weaker learning?
No. Its questionnaires report usage patterns and associations; they do not establish that a particular technology caused score changes. Good policy should evaluate actual learning, privacy, equity and classroom conditions.
What should a principal do first?
Choose a specific learning problem supported by local work samples and teacher judgment, then test a manageable instructional response. Start by observing the quality of student explanations and independent transfer, not by purchasing another platform.
The bigger lesson: protect achievement by strengthening learning itself
The story of PISA 2025 Canada is not that a successful system has suddenly become unsuccessful. It is that a country with substantial educational strengths cannot afford to confuse relative standing with continuous improvement.
The practical response starts with the learner's next move. Can a student notice what they do not understand? Explain a decision using trustworthy evidence? Ask for the right assistance? Carry a strategy from a familiar exercise into an unfamiliar situation?
Those competencies do not fit neatly into a league table. They are also among the qualities that make reading, mathematics and science useful beyond the classroom. Canadian schools can take the PISA warning seriously without treating it as a verdict: teach knowledge thoroughly, assess reasoning thoughtfully, and help students assume more responsibility for learning over time.
For school leaders and teachers who want to explore structured, human-centred learning coaching, discover Learning Coach's professional training pathways and review the specialization most relevant to your work.
References
OECD (8 September 2026), PISA 2025 Results (Volume I): Canada, country note. Official national comparisons, student attitudes, proficiency and data-quality qualification.
Council of Ministers of Education, Canada, PISA 2025. Official Canadian context, assessment scope, provincial reporting and report materials.
OECD, PISA 2025 Assessment and Analytical Framework. Definitions of assessed competencies and domains.
