Just before high school, mathematics lost me at an equation something like:
x = 2 + y
The actual questions were more complicated, but the problem was embarrassingly small.
Make y the subject of the formula.
I could watch somebody move the 2 to the other side and turn it negative.
I could copy the steps.
I could not understand why any of it was allowed.
Then came simultaneous equations, which are not especially friendly to someone who is still privately negotiating with one equation.
Mathematics became hard.
Then it became miserable.
For almost two years, I hated it.
The equation finally meant something
The person who eventually taught me the missing idea was a physics teacher.
Funny enough, he was not trying to teach that topic.
He was explaining velocity:
v = d / t
Velocity is distance divided by time.
That made sense to me. If velocity tells us how much distance is covered per unit of time, then distance must be velocity multiplied by time:
d = v × t
The letters were no longer objects being mysteriously moved across an equals sign.
They represented quantities whose relationship I could picture.
Rearranging the formula was not a typographical ritual. It was another way of stating the same physical relationship.
Once that clicked, I saw the pattern elsewhere.
Equations became easier. Simultaneous equations became possible. The subject I had hated stopped feeling hostile.
By graduation, I was one of the strongest mathematics students in my high-school class.
That ending is satisfying.
The middle bothers me more.
For nearly two years, new mathematics had been placed on top of one missing idea. I had progressed through school, but the prerequisite had progressed with me.
Knowledge has dependencies
School is arranged mostly by age and year.
Knowledge is not.
Some ideas can be learned in almost any order. Others form a dependency graph.
It is difficult to solve simultaneous equations if isolating one variable still feels arbitrary. It is difficult to rearrange a physics formula if the equals sign means “the answer comes next” rather than “these expressions describe the same quantity.” It is difficult to manipulate algebraic expressions if inverse operations are only movements somebody told you to perform.
A simplified mathematics dependency
My story does not prove that every mathematical difficulty comes from one hidden prerequisite.
Research does, however, show the same dependency pattern in another part of mathematics. Eighth- and ninth-grade understanding of fraction magnitude on a number line predicted ninth-grade algebra achievement, even after the researchers accounted for earlier general mathematics achievement and several cognitive measures. (Geary, Hoard & Bailey, 2016)
That does not mean one fraction test can forecast a child’s future.
It means a gap can be structural rather than temporary.
The student is learning two lessons at once
Prior knowledge does more than supply facts.
It changes the mental cost of the new task.
When a knowledgeable student sees two simultaneous equations, they can focus on the new decision:
Should I eliminate a variable or substitute for it?
A student who cannot comfortably rearrange one equation is solving a different problem. They must remember what it means to isolate a variable, which operation undoes another, why the same operation must be applied to both sides and what happens to the sign.
The new lesson on systems of equations is sharing working memory with an old lesson on equivalence.
The student may appear inattentive, slow or careless.
But the task they are performing is not the task the teacher assigned.
They are rebuilding the staircase while trying to climb it.
A passing grade can hide a missing node
Most courses combine many kinds of performance.
A student might earn:
- strong marks on assignments completed with help;
- partial credit for correct procedures;
- high scores on topics that do not depend on the missing concept;
- enough points from participation, projects or corrections;
- a final average just above the threshold.
The grade may accurately say:
Across all evaluated work, this student earned 62 per cent.
It may not answer:
Which prerequisites are secure enough for the next course?
An average permits compensation. Excellent performance in one area can hide a dangerous gap in another.
This is reasonable when the goal is to summarize a course.
It is risky when the next course treats every important prerequisite as present.
But why did the student not learn it?
It is easy to say schools “let students pass,” as if a teacher saw an empty prerequisite and clicked promote out of indifference. Missing knowledge has many histories.
A student may have been absent when the foundation was laid. The class may have needed to move on. Help may exist only after school, when transportation, siblings or paid work make it inaccessible. Knowledge may have faded after the unit test. A total score may reveal that something is wrong without locating it. And asking for help can threaten the identity of a student who believes everyone else learned this years ago.
For me, the earlier explanation simply had not connected. I had seen equations rearranged, but the procedure had not attached itself to a model I could use. The physics teacher supplied that model almost accidentally: velocity, distance and time gave the symbols meaning.
Sometimes the student has not failed to receive an explanation. The explanation has failed to become knowledge.
Resources matter too. We often describe a learning gap as a property of the child:
She is behind.
But “behind” may describe an interaction among time, instruction, opportunity, assessment and support. The child carries the label because the system has nowhere else to put it.
The two blunt options
When the gap becomes visible, schools are often imagined to have two choices.
Retain the student: repeat the grade or course.
Socially promote the student: move them with their age group despite unmet expectations.
Both respond to a real harm. Retention protects the foundation; promotion protects peers, momentum and dignity.
Neither position is foolish.
The evidence does not rescue us with a universal answer.
A 2021 systematic review and meta-analysis examined 84 methodologically sound K–12 studies across multiple countries. On average, retained and non-retained students showed similar development, but results varied by comparison method, country, outcome and time horizon. (Goos, Pipa & Peixoto, 2021)
That average does not mean retention never helps. It means repeating a year is not a reliable general repair mechanism.
That makes sense. If a student repeats the same calendar with the same explanation, time has changed but the instructional diagnosis has not.
Repeating a year is a very large intervention with an uncertain dose of the required treatment. A student who understands most of Grade 8 but cannot rearrange an equation receives another year of everything, different peers and no guarantee that the difficult idea will be taught differently.
Automatic promotion has the opposite mismatch: it preserves peers and momentum but may protect no time for repair. We are choosing between repeating everything and repairing nothing.
I do not wish I had repeated a year. More of the same might only have given me another year to conclude that I was bad at mathematics. I wish the missing idea had been identified earlier and taught differently.
Diagnose → repair → decide
A more useful system would separate three decisions that a final grade currently bundles.
A different progression rule
Diagnose. Use a short, low-stakes assessment that identifies concepts rather than merely ranking students. The result should be:
You understand inverse operations with numbers but do not yet preserve the relationship when the quantities are represented by variables.
Not:
You are a Level 2 learner.
Repair. Give support during the school day where possible, while the student is still learning connected material.
The evidence for tutoring is unusually encouraging. A systematic review and meta-analysis of experimental PreK–12 studies estimated an average effect of 0.37 standard deviations, with stronger average effects for teacher and paraprofessional tutoring and for tutoring delivered during school. It was published as an NBER working paper, so it should not be treated as the final word, but the experimental base is substantial. (Nickow, Oreopoulos & Quan, 2020)
Repair might mean small-group tutoring, a co-requisite support period, a short intensive block, a different representation and a later check that the learning stuck. The important word is funded. “The student should get tutoring” is not a policy if the family must privately purchase it.
Decide. After targeted repair, ask which placement now makes sense. Some students need more time before a highly dependent course; others can progress with support or simply needed the gap named.
Retention remains a possible decision in exceptional cases, especially when missing knowledge is broad and the repeated year will be materially different.
But it should be a designed intervention, not the educational equivalent of turning the device off and on again.
Promotion with an attachment
Perhaps the alternative to retention is not promotion.
It is promotion with an attachment:
This student moves with their cohort, and the school accepts an explicit obligation to repair these named prerequisites.
The plan would name the missing knowledge, the support, the person responsible, the check-in date and what changes if the support is not working. This is less convenient than changing a grade in a database. It also makes the hidden debt visible.
Today, the student often carries that debt privately into the next classroom. The receiving teacher discovers it among many other students. Everyone is surprised by a liability the system itself transferred.
Mastery can become another trap
“Do not move on until every student masters everything” sounds compassionate. It can become rigid. Who defines mastery? Which subjects have true prerequisites? Could endless remediation reduce access to art, music, advanced work or peers?
A dependency-aware system should be strict about the small number of concepts that genuinely unlock later learning and flexible about the path and time used to learn them.
Otherwise mastery becomes one more beautiful word used to defend delay.
Failure has a location
The three questions in this thread now look like one system. High-school grades become difficult to compare. Universities create an exchange rate or let first year perform the calibration. Students arrive carrying whatever was never diagnosed or repaired.
Then we say university is hard. Perhaps it is. But difficulty and surprise are different.
A demanding course can be fair when its prerequisites are clear, its assessments measure the stated capabilities and students have a real chance to prepare.
What feels unfair is discovering that years of successful progression never meant the foundation was secure.
We may congratulate ourselves: no repeated grade, no low final mark, no painful conversation.
But if the missing knowledge simply reappears later—attached to tuition, identity and a transcript—we did not eliminate failure.
We moved it forward.
The humane goal is not to make children fail sooner.
It is to make missing knowledge visible while the repair is still small.
My story ended well, but the repair arrived by accident—from a physics teacher explaining something else. For almost two years, I had interpreted one missing connection as evidence that mathematics itself was beyond me.
A system should not need a lucky collision between the right teacher, the right example and the right afternoon.
And a student eventually becoming excellent does not mean the earlier gap was harmless.
It means the label we were tempted to apply to the student would have been wrong.
Until the next strange question,
Osagie