The Harland Review · The Modern Student

The False Friend

最像的分數,最容易誤讀
For
Parents reading reports from a system not their own
Reading time
13 minutes
Last updated
August 2026
The Reframe

A German parent knows they are translating. A Taipei parent does not.

Picture a familiar scene. A mother in Taipei opens her son's first report from a British school. Her eyes stop on a number. Sixty-eight percent. In Taiwan, her son had never come home below eighty-five, and a strong student is expected in the nineties. Sixty-eight looks like something went badly wrong in a single term abroad.

Nothing went wrong. In a British essay subject, sixty-eight percent is near the ceiling of what the system awards. Seventy percent is the threshold for a first-class mark, and marks above eighty are held back for work approaching publishable quality. One German university's official conversion table states the position flatly: in the British system, seventy percent is already the top mark, and higher marks are given only in exceptional cases. Her son had done well. She was reading a strong result as a near miss.

Now consider why this particular misreading happens, because the mechanism is more interesting than the mistake. A German school report runs from 1 to 6, and 1 is the best grade a child can receive. No parent from Taiwan glances at a German report and assumes they understand it. The scale is visibly foreign, so it announces itself as something to be translated.

The British and American systems do not announce themselves. They use percentages that run in the same direction as Taiwan's, out of the same hundred. Everything about the surface says this needs no translation. That familiarity is the trap. The most expensive misreadings in international education do not happen between systems that look different. They happen between systems that look the same.

Linguists have a name for this failure, and it is exact. They call it a false friend.

What We See
The same family, three systems, three different meanings.

Teaching across the international curricula at once gives an unusual view of this problem, because the confusion arrives in the same household from several directions.

A family may have one child in an American program bringing home A grades. Another sits IGCSEs, where a 7 is a strong result. A cousin works toward a 6 in an IB subject where a 7 is rare. Three scales, three registers, and no shared reference point between them. Parents are asked to hold all of it at once and to know, instantly, which numbers deserve celebration and which deserve a conversation.

The pattern we see is not that parents are careless. It is that they are fluent readers of one system and are handed another that uses the same alphabet. The American A is read as excellence when a great many students are receiving it. The British 68 is read as a shortfall when very few students climb higher. Neither reading is unreasonable. Both are wrong.

The cost is not confusion. Confusion is cheap and self-correcting. The cost is what a parent does next.

A misread number gets acted on. A family buys tutoring their child did not need, against a mark that was already strong. A child who tried hard and did well is met at the kitchen table with visible disappointment, and learns that the standard at home is one no school in that country awards. In the other direction, a family reads a comfortable A and asks no further questions for a year, while the work stays easy. None of those decisions look like errors at the time. All of them start with a number read against the wrong scale.

At a Glance

What a mark is worth

Figures from the awarding bodies and regulators named. Each answers the question a raw number cannot: how many students reach this, and what does the system intend by it.

UK, first-class threshold
70%
Where a first-class mark begins in British higher education, with marks above 80 deliberately rare in essay subjects
UK GCSE, top grades
21.8%
Share of entries at grade 7 or above in 2025, per Ofqual. Grade 9 accounted for 5.1%
UK A-Level, top grade
9.4%
Share of entries awarded A* in 2025, per JCQ. A* and A together reached 27.8%
Taiwan, passing mark
60
The 及格 threshold on the 0 to 100 scale, with 90 and above described as 優秀
IB, average result
30.3
Mean diploma total out of 45 in the May 2024 session, per the IB. Around 9% of candidates reached 40 or above
AP, average score
3.12
Mean score across all AP exams in 2024, per the College Board
US grades today
A
Now the most common grade awarded. At Harvard, 84% of grades were A or A minus in 2024–25
France, excellent
14 / 20
Treated as a strong result. Marks of 16 and above are rare and 20 is effectively never awarded
The Same Number, Twice

Where 68 percent lands

Both systems below run from 0 to 100. Both put the good marks on the right. Nothing on the surface tells a reader that they behave completely differently. Here is the same score read by each.

One score, two systems · percentage scales, identical in appearance
Read as a Taiwanese report
0 to 100, with 60 as the passing mark
不及格 · below passing
及格 to 良好 · pass to good
優秀
68
06090100
68 sits just above the passing line, well short of the range a strong student is expected to reach. Reads as: disappointing.
Read as a British essay mark
0 to 100, with 40 as the passing mark and 70 as first class
fail
pass to upper second
first class · above 80 rare
68
04070100
68 sits at the top of the upper second band, within two marks of first class, in a system where the upper third of the scale is barely used. Reads as: strong.
The number did not move. The scale it was read against did. A parent fluent in one of these systems and handed the other has no visual warning that anything needs translating.

One honest caution belongs here. British grade boundaries are not fixed percentages. Examiners set them each year and by subject, so the raw mark behind a GCSE grade moves. What does not move are the classification thresholds in the essay-marking convention above, and the working assumption behind them, that the top of the scale is reserved, not reached.

Why Similarity Is The Danger

The false friend problem

In language learning, a false friend is a word that looks identical across two languages and means something different. The German word Gift looks like the English gift and means poison. The trap works precisely because the surface is familiar. A learner meeting an obviously foreign word knows to look it up. A learner meeting a word they recognize does not think to check.

Researchers call the wider effect negative transfer, and the finding runs against intuition. Learners make more errors between languages that are closely related, not fewer, because similarity invites them to carry assumptions across unchecked. Distance produces caution. Resemblance produces confidence, and confidence is what allows the error through.

Grades behave the same way. The German 1 to 6 scale is the obviously foreign word: a Taiwanese parent seeing it knows immediately that they cannot read it without help, so they ask. The British percentage is the false friend. It looks like the thing they have read a thousand times. They do not ask, because nothing suggested there was a question.

This is why the parents most at risk of misreading a British or American report are not those least familiar with the West. They are the families whose own system happens to share its shape.

The Scales Themselves

Where the top mark sits

A grade means nothing without two things a number never carries: where the system puts the passing line, and how far up its own scale it is willing to go. Normalized to a common width below, the second of those becomes visible. Several systems reserve their top fifth almost entirely.

United Kingdom, GCSEgrades 1 to 9
pass 4
Grade 7 and above reached by 21.8% of entries, grade 9 by 5.1%
United Kingdom, essay marking0 to 100 percent
pass 40
First class begins at 70. Above 80 is held for work close to publishable
United States0 to 100 percent
pass 60
The scale is used to the ceiling. A is now the most commonly awarded grade
Taiwan0 to 100
及格 60
Used to the ceiling. 90 and above is 優秀, the range expected of a strong student
IB Diplomatotal of 45
pass 24
Mean total 30.3. About 9% of candidates reach 40 or above
Germany6 to 1, inverted
pass 4
Runs backwards, with 1 the best grade, so no reader assumes they understand it
France0 to 20
pass 10
14 is strong. 16 and above is rare and 20 is effectively never given
Netherlands1 to 10
pass 5.5
8 is excellent. 9 and 10 are almost never awarded
Spain0 to 10
pass 5
7 is notable. 9 and 10 are rare
below passing the range in real use top of the scale, rarely or never given
Each scale is drawn to the same width so the proportions can be compared, not the raw numbers. Where a system sets its boundaries annually instead of by fixed percentage, the bands above describe the classification, not the raw score behind it. Sources are listed in full at the end of this piece.
The Words Travel Worse Than The Numbers

What "satisfactory" costs in translation

Numbers at least invite suspicion. The comments beside them rarely do, and they carry more of the meaning.

British school reports are written in a register of understatement that native readers decode without effort and everyone else takes at face value. Sound. Fair. Satisfactory. Working at the expected standard. To a British parent these signal, with varying degrees of politeness, that a child is doing roughly what was asked and no more. To a reader outside that convention they look like praise.

The clearest proof of how badly this travels comes from inside the British system itself. In September 2012 the schools inspectorate abolished the grade satisfactory altogether and replaced it with requires improvement. Schools and parents had been reading satisfactory as acceptable, when it was intended to mean not yet good enough. The Prime Minister of the day put it plainly: just good enough is not good enough. If a single word can mislead native speakers so thoroughly that a regulator retires it, the odds of that word surviving a journey into another language are poor.

Consider where it lands in Chinese. The Taiwanese report vocabulary is graded and precise: 優秀 for excellent, 良好 for good, 及格 for a pass, 待加強 for needs strengthening. Render a British satisfactory into that vocabulary and it settles near 及格, which does not mean adequate in the reassuring British sense. It means the pass you scraped. A comment intended as mild encouragement arrives as a warning.

We should be careful here, because this is our reasoning from the two vocabularies rather than a finding from published research. But the mismatch is not subtle, and any bilingual parent who has read both kinds of report will recognize it immediately.

What The Evidence Will and Will Not Bear

The honest limits of this argument

The case above is strong in places and inferred in others, and the difference is worth stating plainly.

What the evidence supports

The British system compresses its top marks. A first begins at 70 percent and marks above 80 are rare in essay subjects. This is documented in university marking criteria and in official conversion guidance.

Top grades are uncommon. Grade 7 and above at GCSE reached 21.8 percent of entries. A* at A-Level reached 9.4 percent. The IB mean total is 30.3 out of 45.

Similarity worsens transfer. False friends and negative transfer are established findings in language acquisition, and the mechanism carries over cleanly to reading an unfamiliar grading scale.

Where it should stop

Conversion tables exist and are published. Universities, credential evaluators and admissions offices maintain them. The information is available. The failure is that families act on instinct before consulting anything.

Trained evaluators sit at the formal gate. Universities employ people whose job is to read a Taiwanese or British transcript correctly, so the admissions stakes are lower than the anxiety suggests. The damage happens at home, where nobody intercepts it.

Parents adapt. Most families recalibrate within a term or two. The costly window is the first report, before anyone has explained what they are looking at.

Two further honesties. The expectation that a strong Taiwanese student scores in the nineties is well established in the grading vocabulary and in practice. But we could not find a published national distribution to put a figure on it, so we have not invented one. And the argument cuts both ways: the same reasoning that says a British 68 is under-read says an American A is over-read. A parent celebrating a transcript full of A grades in a system where A has become the ordinary grade is making the identical error in the opposite direction.

Common Situations

Four questions worth asking before you react

Not a conversion table. The questions that make a number mean something.

The first report from a new system arrives.
Before reacting, ask the school one question: where does this sit against the year group. A number without a distribution is not information. Most schools will tell you plainly whether a mark is typical, strong, or a concern, and that single answer replaces every conversion table you might otherwise search for.
The mark looks low but the comments are warm.
Trust the comments and check the scale, not the reverse. Teachers writing in a system that compresses marks often signal a strong result in the prose rather than the number. If a British teacher writes that your child is working confidently and contributing well, alongside a mark in the sixties, those two things are not in conflict.
The mark looks high and everything seems fine.
This is the harder case, because nothing prompts you to look. In a system where the top grade has become common, an A tells you your child is somewhere in a large group, not that they are extended. Ask what your child would need to do to be stretched. If the answer is vague, the grade was doing less work than it appeared to.
One word in the report is bothering you.
Ask what it would take to move up one level from that word. The answer converts register into something concrete. A teacher who says satisfactory means secure but not yet independent has told you far more than the word did, and has told you what to work on. Vocabulary varies by system. Specifics do not.
Why We Can Write This Honestly
We read all of these reports, in both languages.

There is an obvious commercial version of this article, and we have tried not to write it. That version tells parents their child's marks are worse than they look, or better than they look, whichever prompts the call. The argument here points in both directions on purpose, because that is what the evidence does.

What lets us write it at all is the vantage. Harland teaches across the American curriculum, IGCSE and A-Level, the IB, and Advanced Placement. Our families mostly grew up inside the Taiwanese system, in a city where a single household routinely holds two or three of these at once. We read these reports weekly, in English and in Chinese, and we have watched the same misreadings recur for years in both directions.

We also know the limits of that vantage. Seeing a pattern in the families we teach is not the same as measuring it, and where this piece runs on reasoning rather than published evidence we have said so instead of dressing inference as fact.

How Harland Helps

Teaching inside every one of these systems, in both languages.

Harland teaches one student at a time across the American curriculum, IGCSE and A-Level, the IB Diploma Programme, and Advanced Placement. Our pedagogy is content-based learning. The practical consequence of working across all of these at once is that we can tell you what a mark means before you decide what to do about it.

01
We can read the report against the right scale.
A mark that looks alarming in one system and a mark that looks reassuring in another often describe the same standard of work. Because our teachers work inside these curricula, not beside them, we can place a result where it belongs before a family commits to a plan built on a misreading.
02
We translate the words, not only the numbers.
Most of what a report communicates sits in its comments, in a register that does not survive translation. Working in English and Chinese with families who move between both, we can tell you what a teacher meant and, more usefully, what the specific next step behind the comment is.
03
We will tell you when nothing is wrong.
An academy that teaches for a living has an obvious incentive to agree that a mark is a problem. Sometimes a 68 is a strong result and the honest answer is that your child is doing well and needs no intervention. We would rather say that and keep your trust than sell a course against a number we knew had been misread.

Not sure what your child's report is telling you?

Bring us the report. We will tell you where the mark sits, what the comments mean, and whether anything needs doing at all. If the answer is that your child is doing well, that is what you will hear.

Start the conversation
PUBLISHED 26 August 2026 · LAST UPDATED 26 August 2026 · All figures verified against the awarding body, regulator, or institution named at the date of publication.
Sources

Where these figures come from

GCSE outcomes at grade 7 and above of 21.8% and grade 9 of 5.1%; A-Level A* of 9.4% and A*–A of 27.8%. Ofqual also confirms that grade boundaries are set after marking each year rather than fixed as percentages.
University marking criteria and conversion guidance, United Kingdom
The first-class threshold of 70% and the rarity of marks above 80% in essay subjects, drawn from published departmental marking criteria and student guidance at UK universities, and from Leibniz University Hannover's official conversion table, which states that 70% is already classed as the top British mark.
Mean diploma total of 30.32 out of 45, mean subject grade of 4.85 out of 7, and the share of candidates reaching 40 points or above.
Mean AP score of 3.12 across all examinations.
National Taiwan University grading regulations and standard Taiwanese grade descriptors
The 0 to 100 scale with 60 as the passing mark, and the descriptor bands 優秀, 良好, 中等, 及格 and 不及格. The expectation that strong students score in the nineties is established in practice and vocabulary; we found no published national distribution and have not implied one.
Ofsted inspection framework change, September 2012
The replacement of the grade satisfactory with requires improvement, following consultation launched in early 2012, on the grounds that satisfactory was being read as acceptable rather than as not yet good enough.
Grade distribution reporting, United States
Evidence that A has become the most commonly awarded grade, including ACT's grade-inflation analysis covering 2010 to 2022, the UCLA CIRP freshman survey series, and institutional reporting that 84% of Harvard grades were A or A minus in 2024–25.
National grading scales, Germany, France, Spain and the Netherlands
Scale directions, passing marks and the rarity of top marks, from official university conversion tables and national study-information services. Verified independently, not taken from any single published comparison chart.
Odlin, Language Transfer (1989), and the research literature on false friends and cross-linguistic influence
The finding that closely related languages produce more negative transfer, which supplies the mechanism this piece applies to grading scales.
World Education Services and UK university international admissions guidance
Credential evaluation practice, including the principle that direct percentage-to-percentage conversion between systems is not valid and that evaluations are country and institution specific.