S

StartSprint Learning Intelligence

Confidence edition

April 2026Monthly Report

The confidence gap: what UK students think they know in April

A calibration analysis of 1,344 student responses across Business, Supply Chain, Science, and Mathematics.

555 sessions·4 teachers·8 subjects

In April, students across StartSprint submitted 1,344 responses across 555 sessions. The headline finding is not the accuracy rate — it is the confidence behind the errors. Across Business, Supply Chain, Science, and Mathematics, students answered incorrectly with high speed and no hesitation on a significant proportion of questions, indicating embedded misconceptions rather than knowledge gaps.

The most dangerous wrong answer is the one a student gives in under two seconds. They are not guessing — they are certain. And certainty is harder to correct than uncertainty.

April's data shows a calibration picture shaped by subject-specific confidence traps: Business and Supply Chain students over-apply simplified mental models to nuanced questions; Mathematics shows the familiar procedural-over-conceptual error pattern; Science students show higher hesitation, which is — paradoxically — a more correctable signal. Where students know they do not know, the teacher's job is easier. Where students believe they know and are wrong, direct intervention is required. This report identifies where each situation is occurring.

Calibration summary

Overconfidence

31%

Wrong answers submitted in under 2s with no changes

Uncertainty

19%

Wrong answers with hesitation or answer changes

Knowledge gaps

9%

IDK selections — students know they don't know

Hesitation delta

+2.9s

Avg time difference between wrong and right answers

Why this matters

April context: with 1,344 responses across only 4 teachers, the per-teacher assessment volume is high — averaging 336 responses each. This means the overconfidence signals in this report are statistically robust. When 31% of wrong answers are given in under two seconds, that is not random noise. It is a consistent pattern of confident wrong mental models embedded across multiple classes and subjects.

Subject breakdown

Business

162 sessions

Cash flow and profit confusion — students apply revenue reasoning to profit questions. Overconfidence is highest in this subject.

Overconfident wrong38%
Uncertain wrong15%
IDK selected7%

Supply Chain

147 sessions

Just-in-time vs just-in-case — students default to JIT as the answer for all inventory questions regardless of context.

Overconfident wrong34%
Uncertain wrong18%
IDK selected8%

Science

118 sessions

Reaction rates and temperature — uncertainty dominates over overconfidence, suggesting students are engaging with the complexity rather than pattern-matching.

Overconfident wrong22%
Uncertain wrong29%
IDK selected12%

Mathematics

94 sessions

Percentage change calculations — students apply additive rather than multiplicative thinking under time pressure.

Overconfident wrong36%
Uncertain wrong21%
IDK selected6%

Economics

14 sessions

Supply and demand shifts — students confuse movement along a curve with a shift in the curve.

Overconfident wrong29%
Uncertain wrong24%
IDK selected11%

Engineering

10 sessions

Force and stress calculations — students substitute area for cross-sectional area inconsistently.

Overconfident wrong27%
Uncertain wrong26%
IDK selected9%

English

6 sessions

Context and purpose identification — students conflate audience and purpose. Limited sample but consistent pattern.

Overconfident wrong25%
Uncertain wrong20%
IDK selected10%

Geography

4 sessions

Development indicators — students over-apply HDI as the universal development measure. Small sample — monitor for April.

Overconfident wrong21%
Uncertain wrong23%
IDK selected14%

Speed signal analysis

Confident wrong1.5s
Fast correct2s
Slow correct5.6s
Uncertain wrong7.1s

Average response time in seconds per answer category. Confident wrong answers are faster than correct ones — the pattern-match is instant.

In the classroom

What April's speed data tells us: confident wrong answers arrived in 1.5 seconds on average — faster than correct answers in many cases. This is the defining signal of an embedded misconception: the brain retrieves a wrong answer instantly because it has been stored and rehearsed. Students who answered slowly and incorrectly (7.1 seconds) were genuinely uncertain — a different teaching problem and a more tractable one. The fast wrong answer requires direct model correction, not additional practice. Practice reinforces the wrong model.

The invisible student

The April invisible student: in Business and Supply Chain, there is a student profile we see repeatedly — high apparent engagement, fast responses, moderate raw score, but a calibration profile that shows 35–40% of wrong answers were given with maximum confidence. This student will perform worse on an exam than their practice scores suggest, because the exam will probe the exact misconceptions they have rehearsed away.

Positive findings

Signal to correction

71%

of teachers who received an April overconfidence signal acted on it within ten days — up from 67% in March. The signal-to-action pipeline is strengthening.

IDK adoption

+14%

increase in IDK selections compared to February across these teachers. Students are becoming more willing to acknowledge uncertainty, which is the diagnostic signal teachers most need.

Top 10 misconceptions

Methodology: each misconception must appear in at least 3 classes and 25 responses in the reporting period. Overconfidence rate = percentage of wrong answers submitted in under 2 seconds with no answer changes. Correct rate = percentage of all attempts answered correctly. April data covers 555 sessions across 4 teachers and 8 subjects.

How to read this list

How to read this list: rank is by overconfidence rate, not by error rate. A question with 60% errors where students hesitate is a content difficulty issue. A question with 40% errors where students answer in 1.4 seconds is a misconception. This list is exclusively about the second type — confident wrong models that will not self-correct without direct intervention.

1
BusinessYear 12·Profitability

A company has revenue of £500,000 and costs of £350,000. What is its profit margin?

8 classes affectedavg 1.4s

Wrong answer most chosen

£150,000

Wrong mental model

Students answer the gross profit figure (£150,000) rather than the profit margin percentage (30%). The question asks for margin — a ratio — but students retrieve the absolute profit value because it is the first calculation step. The fix is explicit emphasis on the distinction between profit (absolute) and profit margin (relative), with consistent use of the % symbol as a trigger to convert.

Overconfidence rate62%
Correct rate31%
2
Supply ChainYear 13·Inventory management

A retailer is concerned about supply disruptions. Which inventory strategy is most appropriate?

7 classes affectedavg 1.3s

Wrong answer most chosen

Just-in-time (JIT)

Wrong mental model

Students select JIT because it is the most rehearsed inventory strategy. The question context — supply disruption concern — explicitly points to just-in-case (JIC) as the appropriate answer. Students are not reading the context; they are pattern-matching to the most salient concept. The intervention is question-reading discipline: "what does this scenario require?" before retrieving a concept.

Overconfidence rate57%
Correct rate34%
3
MathematicsYear 10·Percentage change

A price increases from £80 to £100. What is the percentage change?

6 classes affectedavg 1.6s

Wrong answer most chosen

20%

Wrong mental model

Students compute the absolute difference (20) and treat it as the percentage. The correct calculation divides by the original value: (20/80) × 100 = 25%. The error reflects additive thinking applied to a multiplicative question. Students need the explicit formula anchored: "change over original, not change over new."

Overconfidence rate53%
Correct rate38%
4
BusinessYear 12·Cost classification

Which of the following is a variable cost?

7 classes affectedavg 1.5s

Wrong answer most chosen

Rent

Wrong mental model

Students select rent — a fixed cost — because it is the most prominent cost concept in their working memory. Variable costs vary with output; rent does not. The mental model to correct is the definition anchored to output: "if production stops, does this cost stop? If not, it is fixed."

Overconfidence rate49%
Correct rate43%
5
Supply ChainYear 13·Supply chain fundamentals

What does lead time refer to in supply chain management?

6 classes affectedavg 1.7s

Wrong answer most chosen

The time it takes to sell a product

Wrong mental model

Students confuse lead time (time from order to delivery) with sales cycle time. The word "lead" is interpreted as the front-of-customer process. Explicit definition rehearsal — "lead time is supplier-facing, not customer-facing" — is the correction. Paired contrast with cycle time helps differentiate.

Overconfidence rate46%
Correct rate44%
6
ScienceYear 11·Reaction kinetics

Increasing temperature generally increases reaction rate because:

5 classes affectedavg 2.3s

Wrong answer most chosen

It increases the concentration of reactants

Wrong mental model

Students associate temperature with concentration because both increase reaction rate in their mental model. The correct mechanism is collision frequency and energy: higher temperature means more particles have energy above the activation threshold. Temperature does not change concentration. The intervention should explicitly separate the mechanism (collision energy) from associated but incorrect concepts (concentration).

Overconfidence rate43%
Correct rate47%
7
BusinessYear 12·Stakeholders

Which stakeholder group has the primary legal claim on a company's profits?

5 classes affectedavg 1.8s

Wrong answer most chosen

Employees

Wrong mental model

Students associate value creation with employees and therefore assign profit rights to them. Shareholders hold legal claim on profit distributions (dividends). The confusion reflects a normative moral intuition overriding a legal definition. Clear separation of "who creates value" from "who owns the residual claim" is the intervention.

Overconfidence rate41%
Correct rate49%
8
MathematicsYear 9·Speed, distance, time

A car travels 120 miles in 2.5 hours. What is its average speed?

4 classes affectedavg 1.9s

Wrong answer most chosen

300 mph

Wrong mental model

Students multiply rather than divide, applying the formula inverted. Speed = distance ÷ time (120 ÷ 2.5 = 48 mph). Students who multiply get 300. This is a formula orientation error: the triangle representation is frequently memorised without building the division relationship. Repeated worked examples where "which operation reduces the answer to a sensible number?" is modelled explicitly help correct this.

Overconfidence rate38%
Correct rate51%
9
Supply ChainYear 13·Stock management

What is the main purpose of a buffer stock?

4 classes affectedavg 2.1s

Wrong answer most chosen

To reduce storage costs

Wrong mental model

Students associate stock management with cost reduction — the dominant frame in supply chain teaching — and apply it universally. Buffer stock exists specifically to prevent stockouts when demand or supply varies unexpectedly. Cost reduction is an outcome of optimal stock levels, not the purpose of buffer stock. The intervention is purpose-first teaching: "why does this concept exist?" before "what does it achieve?"

Overconfidence rate35%
Correct rate54%
10
ScienceYear 11·Catalysis

A catalyst increases reaction rate without being consumed. How does it achieve this?

3 classes affectedavg 2.5s

Wrong answer most chosen

It increases the temperature of the reaction

Wrong mental model

Students bundle catalyst and temperature effects because both increase reaction rate and are taught together. A catalyst lowers activation energy; it does not increase temperature. The conflation is understandable given that heating also lowers the effective activation barrier, but the mechanisms are distinct. Visual diagrams of energy profiles — showing activation energy reduction as a dip in the curve — are the most effective correction.

Overconfidence rate33%
Correct rate57%

This month's actions

Urgent

Correct the profit margin vs profit confusion in Business

Business shows the highest overconfidence rate in April (38%), driven by students answering profit absolute values when percentage margin is required. 8 classes affected, average response time 1.4 seconds. Students are retrieving a partially correct answer (the right number, the wrong form) with full confidence.

Recommendation

In the next Business session, build a short routine: before any profitability question, students must identify whether the answer should be in £ or %. Use a StartSprint quiz where all questions make this distinction explicit. The goal is not more calculation practice — it is conditional retrieval: "margin means percentage."

Urgent

Break the JIT default in Supply Chain

Supply Chain shows a persistent pattern: JIT is the most common wrong answer regardless of the question context. 7 classes affected at 57% overconfidence rate. Students have over-learned JIT as the canonical inventory strategy and are applying it indiscriminately.

Recommendation

Create a scenario-matching exercise: give students five supply chain situations and ask them to justify whether JIT or JIC is appropriate. The key intervention is context-reading, not concept recall. Until students read the scenario before selecting an answer, drilling inventory strategy definitions will not fix this.

Watch

Science uncertainty is a healthy signal — do not rush it

Science shows a different profile from Business and Supply Chain: the dominant wrong answer signal is uncertain (29% uncertain wrong vs 22% overconfident wrong). Students are engaging with the difficulty rather than pattern-matching past it. This is a more teachable state — they know there is something they are missing.

Recommendation

Maintain the current pace and continue using StartSprint to track when uncertainty converts to correct confidence versus hardens into overconfidence. The IDK rate of 12% in Science is the highest across all subjects — acknowledge this as healthy metacognition in class to normalise it.