Why your separable equation lost marks

Separation of variables is the workhorse technique of A-Level Mathematics differential equations, and it is also the place where most candidates silently lose two or three marks per question. The idea is mechanical on the surface: rearrange a first-order differential equation of the form dy/dx = f(x)g(y) so that everything involving y sits on one side and everything involving x sits on the other, then integrate both sides. The marks, however, are not awarded for the integration. They are awarded for the algebraic discipline that surrounds it: noticing when the method applies, dividing by the right factor, recovering the constant of integration, and translating the implicit answer into the explicit form the mark scheme wants. This article is a working session on that discipline for A-Level candidates, with an eye on the question types, scoring patterns, and exam format conventions that determine how the method is examined.

The focus throughout is the A-Level specification treatment of separable equations, but the comparison with AP Calculus is useful for students sitting both systems, or for A-Level candidates who want to see where their technique will be tested with more rigour if they move on to university-level work. The same mathematical content is examined, but the mark-scheme emphasis, the tolerance for omitted constants, and the language of the question differ in instructive ways. A solid A-Level grasp of separation of variables is the foundation of a much wider family of differential equation methods: homogeneous equations, linear first-order, and the logistic model all lean on it. Getting it right the first time pays off across the rest of the syllabus.

Recognising a separable equation on the A-Level paper

The first mark for a separation of variables question is awarded, implicitly, before the candidate writes a single line of working. It is the recognition mark. In a typical 6-to-8 mark A-Level question on this topic, the examiner expects the candidate to identify the form, separate the variables, integrate both sides, recover the constant, and apply an initial condition. Each of those stages carries its own cluster of marks, and the recognition step is the hinge: if you misread the form, every subsequent line is wasted. For most candidates reading this, the practical question is not whether separation of variables is on the syllabus, but how the examiner disguises it in the question stem.

The canonical separable form is dy/dx = f(x)g(y), where f depends only on x and g only on y. A-Level questions often present this in disguise. You may see dy/dx = xy, which separates to (1/y) dy = x dx. You may see dy/dx = sin(x)cos²(y), which separates to sec²(y) dy = sin(x) dx. You may even see a model like dN/dt = kN(1 − N/M), the logistic equation, which separates to 1/[N(1 − N/M)] dN = k dt. The pattern is the same in every case: factor the right-hand side into an x-part and a y-part, then split. The recognition skill is to see through a product, a quotient, or a composite function to the underlying separability.

Three diagnostic checks are worth memorising before you write any working. First, can you factor the right-hand side so that one factor contains only x and the other only y? Second, if the right-hand side is a sum, can you rewrite it as a product using a trig identity or an exponent law? Third, if neither of those works, is the equation actually separable, or have you been given a linear first-order equation in disguise? A-Level questions at the higher end of the difficulty range sometimes present an equation that looks separable but is not, and the candidate who charges into separation wastes the first three marks on a method that does not close.

Here is a worked illustration. The equation dy/dx = x²y² + x² looks, at first glance, like it might separate. A candidate who squares y and panics will try to write 1/y² on one side, but the x² term refuses to leave. The correct move is to factor: dy/dx = x²(y² + 1). Now the right-hand side is a product of x² (an x-only factor) and y² + 1 (a y-only factor), and the equation separates cleanly. The mark scheme will give one mark for the recognition, one for the rearrangement, and the rest for the integration and use of the initial condition. Spotting the factor is half the question.

The mechanical step: how to separate and integrate without dropping marks

Once the equation is recognised as separable, the next cluster of marks is awarded for the rearrangement and the integration. This is where A-Level candidates most often lose a single mark in an otherwise correct solution. The mark scheme is unforgiving about notation, division by zero, and the constant of integration. Two working rules cover most of the cases I have marked: write the equation in differential form before you move anything across, and treat the constant of integration as a positive right from the first line, not an afterthought.

The differential form is the cleanest way to separate. Take dy/dx = xy. The first line of working should be (1/y) dy = x dx, written with the differentials on the outside and the functions on the inside. Many candidates write y dy = x dx, missing the negative sign or the reciprocal structure, and lose the first mark. The differential form also makes it obvious when you have made an error: the differentials on the left and right must match the variable in the denominator of the integrand. If you have written y in the numerator on the left, the differential side is dy, so the integrand is y, not 1/y. A 30-second visual check here saves a follow-through error that costs two further marks.

Integration follows. On the left, ∫(1/y) dy = ln|y| + C₁. On the right, ∫x dx = x²/2 + C₂. The two constants collapse into a single constant C, but the way you collapse them is a mark-scheme minefield. A-Level examiners will accept ln|y| = x²/2 + C (with the absolute value bars), or ln y = x²/2 + C provided the domain has been discussed. They will not accept ln y = x²/2 if y could be negative, because the logarithm is not defined for negative real arguments. For most candidates, the safer habit is to write ln|y| every time, even when y is clearly positive from the context. It costs you nothing and pre-empts the mark scheme's pedantry.

The constant of integration should appear on the side where it is most natural to write it. In practice, that is whichever side integrates to a logarithm. The reason is that the constant inside a logarithm can be tidied by exponentiating, and the mark scheme rewards tidy working. A line like ln|y| = x²/2 + C, followed by y = e^(x²/2 + C) = Ae^(x²/2), is the cleanest form. Writing ln y = x²/2 + C on the left and x²/2 + C on the right, then carrying both constants to the right, is technically correct but loses a mark for untidiness in some marking schemes. The preparation strategy here is to write the constant once, on the side that integrates to a log, and to keep it as a single letter until the very last line.

A worked A-Level example will make the pattern concrete. The equation dy/dx = 2xy, with y = 3 when x = 0. Separating gives (1/y) dy = 2x dx. Integrating gives ln|y| = x² + c. Using y = 3, x = 0: ln 3 = 0 + c, so c = ln 3. Then ln y = x² + ln 3, which exponentiates to y = 3e^(x²). Every line is a mark. The separation is one mark, each integration is one mark, the application of the initial condition is one mark, and the final simplification is one mark. Candidates who skip a line by writing y = 3e^(x²) straight from the differential form lose the intermediate working marks, even if the answer is correct.

Initial conditions, constants, and the question types that test them

The use of an initial condition is where A-Level questions on separation of variables diverge most sharply from the corresponding AP Calculus treatment. In AP Calculus, the constant of integration is often absorbed into an arbitrary constant C, and the candidate is expected to evaluate C at the end. In A-Level, the constant is also arbitrary until an initial condition fixes it, but the question types are constructed so that the candidate must explicitly apply that initial condition, and the mark scheme reserves one or two marks for the substitution step. Missing the substitution is the single most common reason a candidate drops from full marks on a separation of variables question.

The standard question type is a 6-mark structured question: state, separate, integrate, substitute, simplify, present. Some A-Level boards extend this to 7 or 8 marks by adding a second part that asks for the value of y at a second value of x, or for the x-intercept of the solution curve. In both cases, the second part is a free ride if the first part has been done correctly, because the candidate already has y as an explicit function of x. Candidates who leave the answer in implicit form, y = e^(x² + ln 3), and then try to evaluate at x = 1 by plugging into the implicit form, often make arithmetic slips that they would have avoided by simplifying first. The mark scheme will accept either form for the evaluation, but the working is safer in explicit form.

A second question type, more common in the harder A-Level papers, asks the candidate to derive a model from a contextual stem and then solve it. The stem might be Newton's law of cooling, an exponential growth model, or the rate of change of a population with a carrying capacity. In each case, the candidate is given a word problem, asked to write down the differential equation, and then expected to separate and solve. The mark allocation is roughly: one mark for the differential equation, three marks for the separation and integration, one mark for the initial condition, and one mark for the interpretation of the final answer in context. The context interpretation is the mark that candidates most often forget. If the question asks for the population after 5 years, the answer 137.4 is worth one mark; 137.4 million, with units, is worth the mark that the examiner is actually testing.

A third question type tests the candidate's understanding of the implicit form, rather than the explicit form, of the solution. The question might ask for the equation of the family of curves satisfying a given differential equation, and accept the answer in the form ln|y| = x²/2 + C, with C as a parameter. This is a different test of the same skill, and the mark scheme often allocates one mark for the constant appearing at all, and one mark for the implicit form being valid for all members of the family. Candidates who collapse the constant too quickly, by writing y = Ae^(x²/2) before being asked, may still receive full credit, but the working is harder for the examiner to follow and occasionally triggers a marking ambiguity. A safe preparation habit is to leave the solution in implicit form until the question explicitly asks for the explicit form.

The silent errors: why your separable equation lost marks even when the answer was right

Separation of variables is unusually prone to silent errors, by which I mean mistakes that do not change the form of the final answer but lose marks for the working. The most common of these, in my experience marking, is dividing by a function of y without checking whether that function is zero. The equation dy/dx = xy, when divided through by y to give (1/y) dy = x dx, implicitly assumes y ≠ 0. The line y = 0 is a singular solution of the original equation, and the candidate who separates without comment has, in the strict reading of the mark scheme, lost the solution y = 0. A-Level boards vary on how heavily they penalise this. Some mark schemes deduct nothing, on the grounds that the singular solution is obvious from the original equation. Others deduct one mark. The safe habit is to add a single line: y = 0 is a solution, and the separation that follows is for y ≠ 0.

A second silent error is the wrong sign in the exponent. The equation dy/dx = 2y, separated as (1/y) dy = 2 dx, integrates to ln|y| = 2x + c, and exponentiates to y = Ae^(2x). A candidate who reads the original equation too quickly and writes 2 dx on the left instead of 2x on the right will get y = Ae^(x²/2) instead, an entirely different function. The mark scheme cannot tell from the final answer which error you made, so it gives the mark for the line where the error occurred. The fix is to write the separation in differential form, with the differentials on the outside, and to read the equation aloud as you do so. Most sign errors become audible.

A third silent error is the use of an indefinite integral on the left and a definite integral on the right, or vice versa, without the limits matching. The equation (1/y) dy = x dx can be integrated from y₀ to y and from x₀ to x, with the limits coming from the initial condition. This form is sometimes called the separation of variables in definite form, and it has the advantage of producing the constant automatically. A-Level questions accept both the indefinite form and the definite form, but a mixed form, with limits on one side and a + C on the other, is treated as a working error. The preparation strategy is to choose one form for a given question and stick to it.

A fourth silent error is the loss of the absolute value on the logarithm. The integrand (1/y) dy integrates to ln|y|, not ln y. The bars matter when y is negative, and they matter in the mark scheme regardless. Candidates who write ln y and then exponentiate to y = e^(x² + C) are technically correct only on the domain where y is positive, and the mark scheme will deduct a mark for the missing absolute value unless the candidate has explicitly noted the restriction. The penalty is small, but across four or five separation of variables questions on a paper, the marks add up to a grade boundary.

Separation of variables in A-Level vs AP Calculus: a comparison that matters for preparation

Students who sit both A-Level Mathematics and AP Calculus AB or BC often assume that separation of variables is the same topic in both systems. The mathematical content is identical. The exam format, the question types, and the scoring are not. Understanding the differences is a preparation advantage, because it tells you which skills to drill harder for which exam. The first difference is the role of the multiple-choice section. AP Calculus has a multiple-choice component in which separation of variables questions appear as one-step or two-step problems, and the candidate is rewarded for picking the right antiderivative form, not for showing working. A-Level, by contrast, is a written examination in which every mark is awarded for working shown on the page.

The second difference is the weight given to the constant of integration. AP Calculus questions often present a separable equation with an initial condition, and the candidate is expected to solve for the constant. The mark scheme, in the free-response section, is similar to A-Level, but in the multiple-choice section the constant is usually provided as part of the answer choices, and the candidate simply has to identify the correct family of curves. A-Level candidates cannot rely on this scaffolding. The constant must be derived, justified, and substituted, and the marks for each step are visible on the page. A preparation strategy that works for AP multiple choice, namely pattern-matching against answer choices, does not transfer directly to A-Level written papers.

The third difference is the role of modelling. AP Calculus BC includes a substantial modelling component, with separation of variables applied to exponential growth, decay, and logistic models. A-Level Mathematics treats modelling as part of the same differential equations topic, but the question types tend to be more structured and the context is shorter. A candidate who has prepared for AP Calculus will find A-Level modelling questions slightly easier in concept but slightly harder in working, because the A-Level mark scheme demands more visible justification of each modelling step. A candidate who has prepared for A-Level will find AP Calculus modelling questions more open-ended, and may need to practise writing the differential equation from a longer word problem.

The fourth difference is the integration repertoire. A-Level integration is built around the standard techniques: substitution, parts, partial fractions, and standard trig integrals. AP Calculus AB stops at u-substitution, and AP Calculus BC adds parts, partial fractions, and trig substitution, but the depth of integration is shallower than A-Level further mathematics. A separation of variables question that integrates to a partial fractions decomposition, 1/[y(1-y)] = 1/y + 1/(1-y), is standard in A-Level and is also standard in AP Calculus BC, but the partial fractions step itself is tested separately in A-Level. The preparation strategy for A-Level is to treat separation of variables as one of several places where partial fractions appears, and to drill the decomposition until it is automatic.

A preparation strategy built around the 5-mark question

The most efficient way to prepare for separation of variables on an A-Level paper is to treat the 5-to-7 mark structured question as the unit of practice, and to drill it until the working is automatic. A single practice session should consist of three to five such questions, drawn from past papers of the relevant board, and timed at 8 to 12 minutes each. The goal of the session is not to get the right answer, but to produce working that the mark scheme would award full marks for. This is a different skill from mathematical correctness, and it is the skill that the A-Level examination format is designed to test.

The first habit to install is the recognition check. Before writing any working, write the equation in the form dy/dx = f(x)g(y) explicitly, and confirm that the factorisation is valid. If the factorisation is not obvious, try a different rearrangement, or use a trig identity. The recognition check is worth one mark on the page, and it is also worth the time you save by not pursuing a non-separable equation down a dead end. A common preparation error is to skip the recognition step because it feels like wasted time, and to dive straight into the integration. The recognition step is not wasted time. It is the only step that prevents the entire solution from being thrown away by a single misread.

The second habit is to write the separation in differential form. (1/y) dy = x dx is a stronger statement than ∫(1/y) dy = ∫x dx, because it shows the candidate has thought about the differentials and not just the integrals. The mark scheme may not explicitly reward this, but it removes a class of working errors that the examiner would otherwise have to detect. The differential form is also a useful self-check: if the differentials do not match the variable in the denominator, the separation is wrong.

The third habit is to apply the initial condition as a substitution step, not as a re-derivation. Once you have ln|y| = x² + c, write the substitution x = 0, y = 3 explicitly, and solve for c. Then substitute c back into the equation. Then simplify. Three lines, three marks. Candidates who try to do the substitution in their head, and write the final answer directly, often lose the substitution mark even when the final answer is correct, because the examiner cannot see the substitution. The mark scheme awards marks for visible working, not for correct answers.

The fourth habit is to leave the answer in the form the question asks for. If the question asks for y in terms of x, give y in terms of x, with the constant of proportionality made explicit. If the question asks for the population after 5 years, give the population after 5 years, with units. If the question asks for the equation of the family of curves, give the family of curves, with C as a parameter. The mark scheme is structured around the form of the answer, and a correct numerical value in the wrong form is treated as incomplete. This is a preparation habit that can be installed in a single practice session, and it pays off across the whole paper.

Common pitfalls and how to avoid them

The pitfalls on separation of variables fall into three families: recognition errors, working errors, and presentation errors. Recognition errors are the most expensive, because they invalidate every subsequent line. The diagnostic check described in the second section, factor the right-hand side into an x-part and a y-part, catches most of them. A specific pitfall is the equation dy/dx = x + y, which is not separable because the right-hand side is a sum, not a product, of an x-term and a y-term. Candidates who try to separate it anyway, by writing dy/(x + y) = dx, will integrate to a log of a sum and end up with a non-closed form. The mark scheme will not award the separation mark, because the method does not lead to a solution within the syllabus.

Working errors are the second family. The most common is the sign error in the exponent, described in the silent errors section. The second most common is the loss of the absolute value on the logarithm. The third most common is the integration of 1/(1 + y) as 1 + y rather than ln|1 + y|. Each of these is a one-mark deduction in a typical mark scheme, and each can be avoided with a 30-second re-read of the working before the candidate moves on. The preparation strategy here is to allocate the last 30 seconds of every practice question to a re-read, and to look specifically for signs, absolute values, and integration results that look suspiciously simple.

Presentation errors are the third family, and they are the most frustrating because they are entirely within the candidate's control. The mark scheme awards marks for visible working, and a candidate who writes the final answer without showing the substitution of the initial condition loses the substitution mark. A candidate who writes ln y without absolute value bars loses the absolute value mark. A candidate who writes 137.4 without units loses the interpretation mark. None of these losses is due to a misunderstanding of the mathematics. They are all due to a mismatch between the candidate's working and the mark scheme's expectations. The fix is a single practice session in which the candidate marks a past paper strictly according to the mark scheme, and then rewrites the solution in the form that would have earned full marks.

A preparation tactic that I have found useful is to keep a log of every separation of variables question attempted, with three columns: the equation, the mark the candidate expected, and the mark the mark scheme would have awarded. The third column is the most informative, because it shows the gap between the candidate's mental model of the working and the examiner's. Over a preparation period of four to six weeks, that gap closes, and the marks stop being lost to presentation. The log takes five minutes per question to maintain, and it is a better return on time than any number of additional practice questions at the same level.

Practising separation of variables under timed conditions

Timed practice is the only way to convert recognition and working habits into exam-day performance. A typical A-Level paper allocates 8 to 12 minutes to a separation of variables question, including the reading time, the working time, and the checking time. Candidates who can solve the question correctly in 15 minutes but not in 10 minutes will run out of time on the paper, and the marks they would have earned will be lost to time pressure. The preparation strategy is to drill the question under timed conditions until the 10-minute target is met consistently, and then to extend the drill to harder questions or to a full paper.

The first stage of timed practice is single-question timing. Choose a separation of variables question from a past paper, set a timer for 10 minutes, and complete the question without referring to notes or the mark scheme. When the timer ends, mark the attempt against the mark scheme, and identify the specific lines that lost marks. The second stage is to repeat the same question the next day, with the goal of reproducing full marks in 10 minutes. The third stage is to attempt a new question of similar difficulty, again under timed conditions. After three to five such cycles, the recognition and working habits are installed, and the timed performance stabilises.

The second stage of timed practice is full-paper timing. A separation of variables question is rarely the only question on an A-Level paper, and the candidate must allocate time across the whole paper. The recommended approach is to budget time per mark, working backwards from the total examination time. A 2-hour paper with 100 marks allows 1.2 minutes per mark. A 6-mark separation of variables question therefore deserves 7.2 minutes. A 7-mark question deserves 8.4 minutes. A 10-mark question, if separation is combined with another technique, deserves 12 minutes. The budget is a guide, not a rule, but it prevents the candidate from over-spending on one question and starving the rest of the paper.

The third stage of timed practice is exam-condition simulation. Sit the paper in a quiet room, with no notes, no phone, and a clock on the wall. Mark the paper under the same conditions that the examiner will use. This stage is psychologically important, because it trains the candidate to manage the stress of the examination and to maintain working standards under pressure. Candidates who have only ever practised single questions in a relaxed environment often find that their working standards slip under exam conditions, and the marks they would have earned in practice are lost in the actual exam. The simulation closes that gap.

For most candidates, four to six weeks of focused timed practice is enough to convert separation of variables from a 4-out-of-6 mark question into a 6-out-of-6 mark question, and to convert a 5-out-of-7 mark question into a 7-out-of-7 mark question. The improvement comes from working on the mark scheme's expectations, not from re-learning the mathematics. The mathematics is already in the candidate's head by the time the preparation period begins. The marks are lost in the working, and the working is what the timed practice installs.

Conclusion and next steps

Separation of variables is a high-yield topic for A-Level Mathematics, both because it appears on virtually every paper and because the marks are concentrated in a small number of working lines that the candidate can control. The preparation strategy is to drill the recognition check, the differential form, the constant of integration, and the initial condition substitution as a single integrated habit, and then to apply that habit under timed conditions. Candidates who treat the question as a 6-mark structured problem, with one mark per line, will find that the marks accumulate quickly and that the topic becomes a reliable source of points on the paper. The next step is a focused practice session on a single 6-mark separation of variables question, marked strictly against the board's mark scheme, and a log of the working lines that lost marks. That diagnostic is the foundation of the preparation strategy.

TestPrep Europe's diagnostic assessment on separation of variables is a natural starting point for candidates building a sharper preparation plan around the 5-to-7 mark structured question.

Frequently asked questions