The Student-to-Teacher Ratio as a Measure of Institutional Quality

1. Introduction

Few metrics in education provoke as much debate as the student-to-teacher ratio. It appears on virtually every school profile, features prominently in accreditation reports, and occupies a central place in the mental calculus of families weighing enrolment decisions. Yet among researchers, the metric's relationship to academic outcomes has been contested for decades, with prominent scholars arriving at conclusions that range from enthusiastic endorsement to cautious scepticism. This tension raises a direct question for any organisation proposing to rank schools on the basis of STR: is the metric defensible?

This paper argues that it is—not because STR is a perfect predictor of educational quality, but because it satisfies a set of structural conditions that are essential for any criterion used in an independent, quantitative ranking system. Specifically, STR is objectively measurable, independently verifiable, resistant to financial manipulation, and directly indicative of the instructional conditions a school provides. No ranking system can rely on a single metric in isolation, but every responsible ranking system must justify each metric it employs. This paper provides that justification for STR in the context of MapleRank's five-criterion ranking model for Canadian private schools.

This paper is written not to persuade leaders that their schools should improve their STR, but rather to explain why STR has been selected as one of five ranking criteria and to demonstrate that the selection rests on a credible evidentiary foundation.

2. Defining the Metric: What STR Measures and What It Does Not

MapleRank defines the student-to-teacher ratio as total enrolled students divided by full-time equivalent (FTE) certified teaching staff. Lower ratios receive higher scores. This definition requires three points of clarification that are important for school leaders considering data submission.

First, the denominator counts only certified teaching staff—individuals whose primary professional function is classroom instruction. Administrators, counsellors, librarians, and teaching assistants who do not hold primary instructional responsibility are excluded. This decision follows the methodological convention used by the Organisation for Economic Co-operation and Development (OECD) in its annual Education at a Glance reports, where the student-teacher ratio is distinguished from the broader student-educator ratio by restricting the denominator to instructional personnel (OECD, 2024).

Second, STR is not the same as class size. As Boozer and Rouse (1995) demonstrated, student-teacher ratio gives an imprecise view of actual class size because teachers may be unevenly distributed across classrooms, some may hold coaching or mentoring roles that reduce their instructional load, and others may team-teach in configurations that inflate apparent headcounts. The OECD has consistently emphasised this distinction, noting that class size and STR "measure very different characteristics of the educational system" (OECD, 2024, Indicator D2). MapleRank selects STR rather than class size precisely because it captures a broader signal: the total human-resource commitment a school makes to instruction relative to its student population, regardless of how that resource is deployed across individual classrooms.

Third, a lower STR does not, by itself, guarantee superior teaching. The research literature is clear on this point. As Hattie (2009) observed in his landmark synthesis of meta-analyses, reducing class size produces only modest achievement effects (an average effect size of d = 0.21) because teachers in smaller settings often fail to adjust their instructional practice to exploit the advantages of fewer students. This finding does not invalidate STR as a ranking criterion; rather, it underscores that STR functions best as one component within a multi-criterion model—which is precisely how MapleRank employs it.

3. The Evidentiary Foundation

The research literature on class size and student-to-teacher ratios spans more than four decades of empirical investigation, encompassing large-scale randomised experiments, international comparative studies, and extensive meta-analytic syntheses. Three bodies of evidence are particularly relevant to the present discussion.

3.1 The Tennessee STAR Experiment

The most rigorous experimental evidence on the effects of reduced student-to-teacher ratios comes from Tennessee's Student/Teacher Achievement Ratio (STAR) project, a four-year randomised longitudinal study initiated in 1985. Approximately 11,600 students across 80 schools were randomly assigned to one of three conditions: small classes of 13 to 17 students, regular classes of 22 to 25 students, or regular classes of 22 to 25 students with a full-time teaching aide (Mosteller, 1995; Word et al., 1990).

The results were unambiguous in direction, if debated in magnitude. Students assigned to smaller classes demonstrated statistically significant improvements in reading and mathematics achievement across all four years of the experiment, with effect sizes ranging from 0.17 to 0.34 standard deviations in favour of the small-class condition (Nye et al., 1993). Critically, the benefits were substantially larger for students from minority and low-socioeconomic-status backgrounds, with initial effects approximately double those observed for majority students (Mosteller, 1995).

The subsequent Lasting Benefits Study confirmed that these gains persisted after students returned to regular-sized classes in later grades. Students who had been in smaller classes in the primary years continued to outperform their peers on achievement measures through at least the eighth grade (Nye et al., 1993). Krueger and Whitmore (2000) extended this analysis further, finding that attending a small class in the early grades was associated with an increased likelihood of taking a college-entrance examination, with the effect being most pronounced among minority students—reducing the Black-white gap in college-test taking by an estimated 54 percent.

"The evidence from Project STAR demonstrated that students randomly assigned to smaller classes showed lasting benefits in academic achievement—benefits that persisted years after they had returned to regular-sized classrooms."
— Mosteller, 1995, The Future of Children

Hanushek (1999) has offered the most prominent critique of the STAR findings, arguing that the commonly cited results may be biased upward due to design and implementation issues and that the effects were limited to very large reductions in class size at the kindergarten level, with no support for more modest reductions or reductions at higher grade levels. This critique is methodologically serious and must be acknowledged. However, even Hanushek's analysis concedes that the directional finding—smaller classes producing positive effects—is not in dispute. The debate concerns magnitude and cost-effectiveness, not the existence of the relationship itself.

3.2 International Comparative Evidence

The OECD's Education at a Glance series provides the most comprehensive cross-national data on student-teacher ratios. The 2024 edition reported that the average STR across OECD countries was 14 students per teacher in primary education and 13 in lower secondary education, with private institutions generally maintaining slightly lower ratios than their public counterparts (OECD, 2024). Between 2013 and 2022, the average number of students per teacher declined by 1.2 across OECD member states, reflecting a broader trend toward increased investment in instructional personnel.

Within Canada specifically, the landscape reveals meaningful variation. Public secondary schools report an average class size of approximately 26.4 students according to PISA data (Allison, 2020), while elite Canadian private schools frequently advertise ratios in the range of 7:1 to 16:1. Upper Canada College, for example, maintains an average class size of approximately 16 students, while the York School in Toronto reports a student-to-teacher ratio of 7:1. These figures are not merely marketing claims; they represent significant investments in instructional capacity that distinguish the private school sector from its public counterpart.

It is worth noting that the relationship between STR and achievement at the aggregate level is complicated by confounding variables. Allison's (2020) analysis of Canadian provincial PISA data found a positive correlation between average class size and achievement scores, with provinces reporting larger classes (such as Quebec and British Columbia) outperforming those with smaller classes (such as Ontario). However, as Allison himself cautioned, this ecological-level correlation does not imply that increasing class sizes would improve individual student outcomes. Provincial achievement differences reflect the cumulative influence of curriculum design, teacher training, school culture, and socioeconomic composition—variables that operate far upstream from the point at which a student sits in a classroom of a given size.

3.3 Meta-Analytic Syntheses

Hattie's (2009) Visible Learning synthesis of over 800 meta-analyses assigned class-size reduction an average effect size of d = 0.21, ranking it 132nd among 138 educational influences studied. This finding has been widely cited by opponents of class-size reduction as evidence that the practice is not worthwhile. However, the interpretation requires substantial qualification.

Glass and Smith (1979), whose meta-analysis was among those synthesised by Hattie, had reported considerably larger effects for substantial reductions in class size (from 25 or more to 15 or fewer students), particularly in well-controlled studies. Hagemeister (2020) and others have argued that Hattie's method of averaging across all class-size reductions—including trivially small reductions that would not be expected to produce meaningful effects—diluted the signal from studies examining educationally significant reductions. Nielsen and Klitmøller (2017) further observed that the meta-analyses combined in Hattie's synthesis used inconsistent definitions of both "small" and "regular" classes, introducing definitional noise that would tend to attenuate observed effects.

More recent meta-analytic work has provided additional nuance. Research on teacher-student relationships has consistently found that the affective quality of these relationships is significantly associated with both student engagement and achievement. Roorda et al. (2011), in a meta-analysis of 99 studies encompassing over 88,000 students, found that positive teacher-student relationships were strongly associated with engagement and modestly associated with achievement across all grade levels from preschool to high school. While this finding does not directly address class size, it establishes a clear mechanistic pathway through which lower student-to-teacher ratios could enhance educational outcomes: fewer students per teacher creates the structural conditions under which closer, more sustained teacher-student relationships become feasible.

4. Beyond Achievement Scores: The Broader Case for STR

The debate about STR has been disproportionately dominated by its relationship to standardised test scores. This framing, while understandable given the methodological demands of quantitative research, obscures a set of educational values that are particularly salient in the Canadian private school context and especially relevant to the international families whom these schools serve.

4.1 Socio-Emotional Development and Pastoral Care

International families investing $200,000 to $300,000 in a Canadian private school education are purchasing far more than test preparation. They are entrusting their children—often at age 14 or 15, often across linguistic and cultural boundaries, and often for the first time living away from home—to an institution that will serve as the primary community in which those children mature during the most formative years of their social and emotional development.

In this context, the student-to-teacher ratio takes on significance that transcends its relationship to academic achievement. Research on socio-emotional development in secondary school settings has consistently identified the teacher-student relationship as a key protective factor for adolescent well-being. Positive relationships with teachers are associated with enhanced self-esteem, reduced depressive symptoms, greater school belonging, and improved behavioural adjustment across diverse cultural contexts (Jia et al., 2009; Roorda et al., 2011). These effects are not trivial considerations for a boarding student from Shenzhen or Dubai navigating acculturation, language barriers, and separation from family.

Lower student-to-teacher ratios do not automatically produce these relational benefits, but they create the structural preconditions under which such relationships become possible. A teacher responsible for 120 students across five sections per day operates under fundamentally different relational constraints than a teacher responsible for 60. The former may know students' names and grades; the latter has the capacity to know their aspirations, anxieties, and the texture of their adjustment to a new country.

4.2 Differentiated Instruction and English-Language Support

The Canadian private school sector serves an increasingly diverse international student body. Schools that actively recruit from multiple countries and language backgrounds face a specific instructional challenge: providing rigorous academic content to students whose English-language proficiency may vary dramatically from native fluency to intermediate competence. Effective differentiated instruction—the practice of adjusting content, process, and product to meet individual learner needs—is pedagogically demanding and practically constrained by the number of students a teacher must simultaneously serve.

Lower student-to-teacher ratios provide teachers with the time and cognitive bandwidth to monitor individual comprehension, offer targeted feedback, and adapt instruction in real time. These are precisely the instructional behaviours that Hattie (2009) himself identified as having the highest effect sizes in his synthesis: feedback (d = 0.73), formative evaluation (d = 0.90), and teacher clarity (d = 0.75). The paradox of Hattie's findings on class size is that the instructional practices most strongly associated with achievement gains are also the practices most constrained by high student-to-teacher ratios. A school cannot deliver high-quality individualised feedback to 30 students with the same frequency and depth as it can to 15.

"The instructional practices most strongly associated with student achievement—individualised feedback, formative assessment, teacher clarity—are precisely those whose quality degrades most rapidly as the number of students per teacher increases."

4.3 Safeguarding and Duty of Care

For international boarding students, the school functions in loco parentis to a degree that extends well beyond instructional hours. Adequate student-to-teacher ratios are a prerequisite for the observation, monitoring, and relational proximity that effective safeguarding requires. A school's capacity to notice when a student is withdrawn, struggling socially, or at risk is directly constrained by the number of students for whom each adult is responsible. While safeguarding protocols and training are essential, they operate within structural parameters that STR quantifies.

5. Why STR Satisfies the Conditions of a Sound Ranking Criterion

MapleRank's ranking methodology is built on a foundational principle: each criterion must be selected not because it is a perfect measure of school quality, but because it satisfies a set of structural conditions that make it defensible for use in an independent, quantitative ranking. The student-to-teacher ratio satisfies four such conditions.

5.1 Objective Measurability

STR is calculated from two inputs: total student enrolment and total FTE certified teaching staff. Both figures are unambiguous, countable, and expressible as a single number. There is no subjective judgment involved in the calculation, no qualitative assessment, and no evaluator bias. This objectivity is not a trivial property. Many plausible indicators of school quality—teaching effectiveness, school culture, leadership quality—resist quantification in ways that make them unsuitable for inclusion in a standardised ranking. STR does not suffer from this limitation.

5.2 Independent Verifiability

The inputs to the STR calculation can be cross-referenced against publicly available data, accreditation records, regulatory filings, and employment records. A school reporting an implausibly low STR can be flagged for verification against external sources. This verifiability is essential for a ranking system whose credibility depends on the integrity of the data it processes. MapleRank's four-phase verification protocol—self-reported data, spot-checks by the Chief Academic Officer, third-party audits, and AI-driven anomaly detection—provides layered assurance that reported STR figures reflect institutional reality.

5.3 Resistance to Financial Manipulation

A school cannot improve its STR through marketing expenditure, agency relationships, or payment to the ranking platform. The only way to improve a school's STR is to hire additional qualified teaching staff or reduce student enrolment—both of which represent genuine institutional investments in instructional capacity. This structural resistance to gaming is the cornerstone of MapleRank's editorial independence. Unlike qualitative reputation surveys, which can be influenced by brand recognition and marketing spend, STR rewards actual resource allocation.

5.4 Meaningful Association with Educational Conditions

As the preceding sections have demonstrated, STR is not merely a financial accounting metric. It captures the human-resource infrastructure within which instruction, pastoral care, differentiated support, and safeguarding occur. The relationship between STR and student outcomes is not simple, not linear, and not deterministic—but it is real, directionally consistent, and supported by the most rigorous experimental evidence available in education research.

6. Addressing Legitimate Objections

Institutional leaders considering participation in MapleRank's ranking system may harbour reasonable concerns about the use of STR as a criterion. This section addresses the three most common objections.

6.1 "STR Does Not Capture Teaching Quality"

This objection is valid on its own terms. A school with a 10:1 STR staffed by ineffective teachers will not produce better outcomes than a school with a 15:1 STR staffed by excellent ones. However, this objection applies equally to every quantitative metric in education: IB scores do not capture the development of character, university matriculation rates do not measure intellectual curiosity, and diversity indices do not assess the depth of cross-cultural understanding. The purpose of a quantitative ranking criterion is not to capture everything that matters, but to capture something that matters in a way that is objective, verifiable, and comparable. STR meets this standard. Teaching quality is addressed through a separate criterion—faculty subject expertise—within MapleRank's multi-criterion model.

6.2 "A School Could Reduce STR by Hiring Unqualified Staff"

MapleRank's definition explicitly restricts the denominator to certified teaching staff. Teaching assistants, administrative personnel, and uncertified instructors are excluded. A school that hired uncertified individuals solely to improve its reported STR would see no benefit in the ranking calculation. Additionally, MapleRank's verification protocol includes cross-referencing reported staffing data against publicly available information and regulatory records, providing a further check against inflation.

6.3 "Our STR Is Higher Because We Invest in Other Areas"

This is a legitimate institutional choice, and MapleRank's ranking model accommodates it. Because STR is one of five weighted criteria, a school with a higher STR can compensate through stronger performance on other dimensions—superior university matriculation rates, higher IB scores, greater student body diversity, or more extensively qualified faculty. The ranking rewards overall institutional quality, not performance on any single metric. A school that allocates resources toward laboratory facilities rather than additional teachers may score lower on STR but higher on university matriculation outcomes if those facilities contribute to stronger student preparation. The multi-criterion structure is designed to honour precisely this kind of institutional variation.

7. STR in the International Decision-Making Context

The international families whom Canadian private schools recruit operate within a decision-making framework that is fundamentally different from that of domestic families choosing between a local public school and a nearby independent alternative. An international family is making a commitment that typically exceeds $200,000 over four years, involves sending a minor child to live in a foreign country, and is often informed by limited and unreliable information. For these families, the student-to-teacher ratio carries a specific and powerful signal.

In markets such as China, South Korea, and the Gulf States, class size is widely understood as a marker of educational investment. Large class sizes are associated with public-sector mass education; small class sizes signal premium, individualised instruction. Whether or not this association is fully supported by the nuances of educational research, it reflects a deeply embedded consumer expectation in the markets from which Canadian private schools draw their international enrolment. A school that ranks well on STR communicates to these families that their tuition investment translates into tangible instructional capacity—that their child will not be lost in a crowd.

This market-signalling function does not, by itself, justify the inclusion of STR in a ranking model. MapleRank's criteria are selected on the basis of educational defensibility, not consumer preference. But the convergence of educational evidence and market expectation in the case of STR is notable. The metric satisfies both the analytical requirements of a credible ranking and the informational needs of the families the ranking is designed to serve.

8. How MapleRank Operationalises STR

Schools participating in MapleRank's ranking submit two data points for the STR criterion: total current student enrolment and total FTE certified teaching staff. The ratio is calculated by MapleRank's algorithm and scored on a continuous scale in which lower ratios receive higher scores. The raw data is treated as confidential and is never displayed publicly; only the calculated ranking position and aggregate score are visible.

Schools that do not submit data for any criterion remain unranked. MapleRank's verification protocol applies to all submitted data, including STR, through a phased approach: self-reported submission, Chief Academic Officer spot-checks, third-party audits as the platform matures, and continuous AI-driven anomaly detection.

The weighting assigned to STR within the five-criterion model is proprietary and not disclosed. This deliberate opacity prevents schools from optimising for a single metric at the expense of holistic institutional quality. Schools are informed of the five criteria and their definitions but must invest across all dimensions to maximise their ranking position.

9. Conclusion

The student-to-teacher ratio is neither a perfect measure of school quality nor a sufficient one. No responsible ranking system would rely on it in isolation, and MapleRank does not. What STR provides is a structurally sound, objectively measurable, independently verifiable, and educationally meaningful indicator of the human-resource commitment a school makes to instruction, pastoral care, and the conditions under which learning occurs.

The experimental evidence from Project STAR demonstrates that lower student-to-teacher ratios produce directionally consistent benefits for student achievement, with particularly strong effects for students from underrepresented backgrounds. The international comparative data from the OECD confirms that STR is a standard metric used worldwide to assess educational resourcing. And the broader research literature on teacher-student relationships, socio-emotional development, and differentiated instruction establishes clear mechanistic pathways through which lower ratios enhance the quality of the educational experience.

For Canadian private schools serving international families, STR carries additional significance. It communicates a commitment to individualised attention that aligns with the expectations of families investing hundreds of thousands of dollars in their children's education. It signals capacity for the pastoral care and safeguarding that boarding students require. And it provides a quantitative anchor for claims of small class sizes and personalised instruction that every school makes but few can independently substantiate.

MapleRank invites school leaders to engage with the ranking process not as a threat to institutional autonomy but as an opportunity to demonstrate, through independently verified data, the investments they have made in instructional quality. The student-to-teacher ratio is one dimension of that demonstration—a necessary one, grounded in evidence, and deployed within a multi-criterion framework that honours the complexity of what it means to be a great school.

References

Allison, D. J. (2020). Secondary school class sizes and student performance in Canada. Fraser Institute. fraserinstitute.org

Boozer, M. A., & Rouse, C. E. (1995). Intraschool variation in class size: Patterns and implications. Journal of Urban Economics, 50(1), 163–189. doi.org

Durlak, J. A., Weissberg, R. P., Dymnicki, A. B., Taylor, R. D., & Schellinger, K. B. (2011). The impact of enhancing students' social and emotional learning. Child Development, 82(1), 405–432. doi.org

Finn, J. D., & Achilles, C. M. (1990). Answers and questions about class size. American Educational Research Journal, 27(3), 557–577. doi.org

Glass, G. V., & Smith, M. L. (1979). Meta-analysis of research on class size and achievement. Educational Evaluation and Policy Analysis, 1(1), 2–16. doi.org

Hagemeister, V. (2020). A critical analysis of the section on class size in Hattie's Visible Learning. ResearchGate

Hanushek, E. A. (1999). Some findings from an independent investigation of the Tennessee STAR experiment. Educational Evaluation and Policy Analysis, 21(2), 143–163. doi.org

Hattie, J. (2009). Visible learning: A synthesis of over 800 meta-analyses relating to achievement. Routledge.

Jia, Y., et al. (2009). The influence of student perceptions of school climate on socioemotional and academic adjustment. Child Development, 80(5), 1514–1530. doi.org

Krueger, A. B., & Whitmore, D. M. (2000). The effect of attending a small class in the early grades on college-test taking and middle school test results. NBER Working Paper No. 7656.

Mosteller, F. (1995). The Tennessee study of class size in the early school grades. The Future of Children, 5(2), 113–127. doi.org

Nielsen, K., & Klitmøller, J. (2017). Blind spots in Visible Learning. Journal of Academic Perspectives, 2017(1), 1–22.

Nye, B. A., Zaharias, J. B., Fulton, B. D., & Achilles, C. M. (1993). The lasting benefits study. Center of Excellence for Research in Basic Skills, Tennessee State University.

Organisation for Economic Co-operation and Development. (2024). Education at a glance 2024: OECD indicators. OECD Publishing. oecd.org

Roorda, D. L., Koomen, H. M. Y., Spilt, J. L., & Oort, F. J. (2011). The influence of affective teacher–student relationships on students' school engagement and achievement. Review of Educational Research, 81(4), 493–529. doi.org

Word, E., et al. (1990). Student/Teacher Achievement Ratio (STAR): Tennessee's K–3 class size study: Final summary report, 1985–1990. Tennessee State Department of Education.