Separable versus linear ODEs on the Calculus exam

AP Calculus students often meet the words 'differential equation' for the first time in the autumn of their final A-Level-equivalent year and immediately treat it as a separate chapter bolted onto the end of the syllabus. In practice, differential equations are simply the language that the rest of calculus has been building toward: derivative, integral, rate of change, accumulated area, all joined into a single sentence that says 'here is how something evolves'. On the AP Calculus AB and BC papers, the topic carries its own weight in Unit 7 (Differential Equations) and reappears inside Units 4, 5, 6, 8 and 9 wherever rates, accumulation, or modelling are tested. A candidate who walks into the exam with a clean, narrow understanding of the half-dozen families that actually appear can pick up a substantial block of marks without ever invoking anything exotic.

For learners working inside an A-Level preparation strategy that overlaps with AP-style study, the right frame is straightforward. Treat the topic as three linked skills: write an equation from a verbal description, solve it using a small menu of techniques, and interpret the solution in the context of the original problem. Everything in the unit reduces to those three moves. The pages that follow walk through the families, the techniques, the question types, and the scoring logic that mark schemes use when awarding credit. The piece is written for students aiming at the 4 to 5 grade band on AP and at A* at A-Level, but the modelling instinct that sits underneath is the same in either specification.

What the exam actually means by a differential equation

A differential equation is an equation that relates a function to one or more of its derivatives. On the AP Calculus paper, the form you meet almost without exception is dy/dx = f(x, y), where the right-hand side is some expression in x and y that the paper has already given you. The candidate's job is to recognise the family, apply the right technique, and produce a solution in a form the rubric recognises. That recognition is the test of whether you have understood the unit, and the recognition is shallow only on the surface. Underneath it, there are six or seven predictable families, and once a student can spot the family in five seconds, the rest of the work follows.

The reason the exam uses this topic is that it forces every other calculus skill to appear in disguise. To solve dy/dx = 3x²y the student has to separate variables, integrate a polynomial, recover the constant of integration, and read the resulting exponential against the original context. Every one of those micro-skills is reused, and a single careless step can cost credit on a problem that, at heart, only asked for the constant. Markers know this. Most rubric items in this unit are partitioned so that a candidate can earn at least one of the two or three points available even if a later step collapses. Preparation strategy for the topic should aim at that pattern: a full marks attempt on the separable equation, a defensible statement of the slope field, and a careful interpretation of the constant.

Verbal stems the exam reuses

Three sentence templates cover the majority of stems on the topic. First, 'the rate of change of y is proportional to y', which the paper then writes as dy/dx = ky and expects a candidate to integrate to a base-e exponential. Second, 'the rate of change of y is proportional to the difference between y and a constant', which gives a linear first-order ODE whose general solution is a horizontal asymptote at that constant. Third, 'a tank contains ... grams of salt in ... litres of water, brine enters at ... and leaves at ...', which is the standard mixture problem and which reduces to dA/dt = rate in - rate out with the second term carrying a division by the volume. If a candidate can pattern-match a stem to one of these in a single reading, the question is half-finished before any algebra begins.

Separable equations: the most reliable marks on the topic

A separable differential equation is one that can be rewritten so that all of the y terms (including dy) sit on one side and all of the x terms (including dx) sit on the other. The exam almost always uses the form dy/dx = f(x)g(y), and the standard opening move is to divide through by g(y) and multiply through by dx. Two examples illustrate the pattern that the exam favours. Given dy/dx = 2xy, divide both sides by y and integrate: the left side gives ln|y| = x² + C, the right side integrates to x² + C, and the solution rearranges to y = ±e^{C} e^{x²}, which the rubric accepts as y = C e^{x²}. Given dy/dx = x/y, separation gives y dy = x dx, integration gives y²/2 = x²/2 + C, and the implicit solution is acceptable as the final form.

Two tactical details decide whether a candidate keeps the marks. The first is the constant of integration. AP rubrics require a '+ C' on the right side of the integrated equation, and a candidate who omits it loses one point even when the algebra is otherwise perfect. The second is the domain restriction that follows from dividing by a function of y. A solution such as y = C e^{x²} is valid only where the original separation was defined, so for dy/dx = 2xy the solution excludes y = 0, and a candidate who treats y = 0 as a valid solution without justification will be marked down. The same logic applies to dy/dx = 1/(y-3), where the rubric expects the student to note that y = 3 is excluded and that the constant of integration carries a sign through the logarithm.

How the marker awards credit in two or three points

On a typical AB question, the rubric is partitioned into three steps. The first step is the separation itself, worth one point, and the marker is checking that all of the y terms sit on one side and all of the x terms on the other. The second step is the integration, worth one point, with credit awarded only when both sides are integrated correctly. The third step is the solution in acceptable form, worth one point, and the rubric usually accepts either an explicit y = f(x) or an implicit F(x, y) = C. Knowing this partition is itself a preparation strategy, because it tells the candidate which mistakes are recoverable: a slipped sign on the second integral is recoverable if the separation step is recorded, while a missed separation is not. Most candidates reading this who have lost points on the topic lost them on step one, not step three.

Slope fields, Euler's method, and the reading marks they earn

Slope fields are a graphical representation of a differential equation: at a grid of points (x, y), the marker draws a short line whose slope is given by the right-hand side f(x, y). The exam asks three things of a candidate. First, given a small slope field, the candidate may need to identify which differential equation produced it, and the test is essentially pattern recognition: the slope along the y-axis, the slope along a horizontal line, and the slope along a vertical line. Second, the candidate may need to sketch the solution curve that passes through a given initial condition, drawing a smooth curve that follows the printed line segments as closely as possible. Third, on a multiple-choice item, the candidate may be asked which of four curves is consistent with the field, and the wrong answers are usually the curve with the wrong asymptotic behaviour or the curve that crosses an obvious isocline.

Euler's method is the computational partner of the slope field. Given a step size h and a starting point (x₀, y₀), the candidate computes y_{n+1} = y_n + h f(x_n, y_n), recording each pair of values in a table. The exam asks for the next point after one or two iterations, and a candidate who has lost the method on the day can usually recover it from a single example worked in a textbook. The tactical mistake to avoid is treating the method as exact: Euler's method underestimates or overestimates depending on the curvature, and a candidate who reports the value to three decimal places when the rubric only expects an answer accurate to the first decimal will not lose the mark, while a candidate who rounds too aggressively to a single decimal may. The exam does not penalise a candidate for writing more steps than were strictly necessary, only for fewer.

Common pitfalls in slope-field and Euler items

The most expensive mistake in slope-field items is to assume that the curve must pass through a grid point, and the second most expensive is to ignore the field at the boundary. In Euler's method, the most expensive mistake is sign error in the slope, particularly when the right-hand side contains a subtraction. A candidate who sets up the table with the wrong sign on the increment loses all subsequent iterations and forfeits the trajectory-shaped mark. The preparation strategy is simple: write the table with the function f written out as an expression in x and y, plug in the numbers, and only then perform the arithmetic. The expression does not need to be simplified; the marker accepts a numerical answer and an entry in the table that shows the work.

Exponential growth and decay: the four patterns the marker recycles

The exponential family is the most heavily tested sub-topic in the differential-equations unit, and it almost always appears in one of four guises. The first is unconstrained growth, where the rate of change is proportional to the quantity itself: dP/dt = kP, with k > 0 giving growth and k < 0 giving decay. The solution is P(t) = P₀ e^{kt}, and the exam usually gives the initial value in a sentence and asks the candidate to find the constant. The second is Newton's law of cooling, where the rate is proportional to the difference between the object and its surroundings: dT/dt = k(T - T_s), with the solution asymptoting to T_s as t grows. The third is logistic growth, where the rate is proportional to both the population and the remaining capacity: dP/dt = kP(1 - P/M), with the solution carrying a horizontal asymptote at M. The fourth is radioactive decay, which is just unconstrained decay with a negative k and a half-life translation: T(t) = T₀ e^{-λt}, where λ = ln 2 / t_{1/2}.

Each pattern has a characteristic use of the constant. In Newton's law of cooling, the constant is sometimes expressed in terms of the half-life of the temperature gap, and a candidate who reports the time at which the gap halves is working with the right structure even if the algebra goes sideways. In logistic growth, the constant of integration is absorbed into the (M - P) factor and the explicit solution is a fraction rather than an exponential: P(t) = M / (1 + A e^{-kt}), with A determined by the initial condition. In radioactive decay, the half-life language lets the candidate avoid the explicit logarithm: a problem that says 'after 12 years only a third of the sample remains' should be read as a half-life problem, and the candidate should write down that the constant is -(ln 3) / 12 before any further work. Candidates who skip this step and go straight to the integration of a separable equation end up solving a harder version of the same problem and typically lose the half-life point.

Exponential family: a worked micro-example

Consider the question 'A bacteria culture grows at a rate proportional to its size. After two hours the population is 1500 and after five hours it is 4500. Find the population at t = 0.' The structure is unconstrained growth, the differential equation is dP/dt = kP, and the separation gives ln P = kt + C or P = P₀ e^{kt} in explicit form. The candidate takes the ratio of the two observations: 4500 / 1500 = e^{3k}, so 3k = ln 3 and k = (ln 3) / 3 per hour. Back-substituting, 1500 = P₀ e^{2k} gives P₀ = 1500 / e^{2k} = 1500 / e^{2 ln 3 / 3} = 1500 / 3^{2/3}. Numerically this is roughly 830. The full solution earns the integration point, the constant point, and the answer point on a typical rubric. A candidate who reports a different number because they divided instead of multiplying when taking the ratio still receives the integration and the constant points, which is the second reason to record every step on the page.

Linear first-order equations and the integrating factor

Linear first-order equations are the second most common family and they appear most often in the standard form dy/dx + P(x)y = Q(x), where P and Q are functions of x alone. The standard solution technique is the integrating factor, defined as μ(x) = e^{∫P(x) dx}. Multiplying both sides of the standard form by μ(x) produces an exact derivative: d/dx[μ(x) y] = μ(x) Q(x), and the right side can be integrated directly. The exam almost always uses an integrating factor that is a single exponential, and the most common patterns are P(x) = 1 giving μ = e^{x}, P(x) = 2x giving μ = e^{x²}, and P(x) = -k giving μ = e^{-kx}. The candidate who can integrate a polynomial against such a μ has solved the problem; the rest is bookkeeping.

The BC paper, in particular, has at least one item per sitting that tests the integrating factor in some form. On the AB paper the topic is sometimes optional or partially tested, but the structure of the technique reappears inside the exponential decay family, where the differential equation is also linear. A candidate who treats every decay problem as a separable equation is doing extra work but is not doing wrong work, and the rubric will usually accept either path. The integrating factor is the cleaner path when the right-hand side is not a pure multiple of the dependent variable.

Worked micro-example: integrating factor

Consider dy/dx + 2y = x. The integrating factor is μ(x) = e^{∫2 dx} = e^{2x}. Multiplying through gives e^{2x} dy/dx + 2 e^{2x} y = x e^{2x}, which the candidate recognises as d/dx[e^{2x} y] = x e^{2x}. Integrating the right side by parts or by table gives e^{2x} y = (x/2 - 1/4) e^{2x} + C, and the solution is y = x/2 - 1/4 + C e^{-2x}. The constant of integration is determined from the initial condition if one is given. The integration step itself can be verified by differentiation: dy/dx = 1/2 - 2C e^{-2x}, and the right side of the original equation is x - 2y = x - (x/2 - 1/4 + C e^{-2x}) = x/2 + 1/4 - C e^{-2x}, which differs by a sign and a constant; the candidate should re-check the integration when this happens, and the marker will usually accept a corrected attempt written clearly above the previous line. This is a habit, not a fact: a second attempt scored cleanly is worth as much as a first attempt that was always right.

Logistic and restricted growth: reading the asymptote

Logistic growth appears in both AB and BC items, but the form on the AB paper is more often a verbal problem that the candidate is expected to recognise and write down, while the BC paper asks the candidate to derive or interpret the general solution. The differential equation dP/dt = kP(1 - P/M) is separable, and the separation gives a partial-fraction integral of the form ∫ dP / [P(1 - P/M)] = ∫ k dt. The standard move is to write the left side as a sum of two simpler fractions and integrate each: (1/P) + (1/(M-P)) = M / [P(M-P)], so that ∫ M / [P(M-P)] dP = (1/P + 1/(M-P)) dP / 1, and the integral on the left becomes ln|P| - ln|M-P| = kt + C. The solution in explicit form is P(t) = M / (1 + A e^{-kt}), where A depends on the initial population.

The exam asks the candidate to read three things off the solution. The first is the initial value, used to compute A. The second is the long-term behaviour: as t grows, the exponential term vanishes and P(t) → M, so M is the carrying capacity. The third is the inflection point, which on a logistic curve occurs at P = M/2, and the candidate who can identify the inflection point from the differential equation itself, by setting the second derivative to zero, has the structure of the curve. The error pattern to avoid is the sign in the exponential: a candidate who writes P(t) = M / (1 + A e^{kt}) has a population that decays to zero rather than grows to M, and that mistake is not recoverable by a sign on A alone.

Tactical notes on carrying capacity

The most common verbal form of the logistic equation in the AP exam is 'a population grows at a rate jointly proportional to the population and to the difference between the carrying capacity and the population'. The candidate who reads this sentence and writes dP/dt = kP(M - P) has the differential equation and can earn the first mark. The second mark is the separation, the third is the integration, and the fourth is the explicit form. A candidate who cannot complete the integration in the heat of an exam should still write the differential equation and the separation, because both earn credit on a typical rubric. This is the tactical backbone of the unit: never go to the page without writing the first line, because the first line is what the rubric scores first.

Mixture problems, tank models, and the language of concentration

Mixture problems are the most verbal of the differential-equation items, and they sit at the intersection of the calculus and the model. The standard setup is a tank of brine: an inflow pipe brings a saline solution with a known concentration at a known rate, an outflow pipe removes the mixture at a known rate, and the candidate is asked to find the amount of salt in the tank at a particular time. The differential equation is dA/dt = (concentration in)(rate in) - (A/V)(rate out), where A is the amount of salt, V is the volume, and the second term carries the concentration of the tank itself. If the volume is constant, the equation is dA/dt = r c_in - (A/V) r_out and the variable on the right is A alone, making the equation linear and separable.

The exam is friendly to the candidate on this item type, because the differential equation is essentially a balance equation that any first-year chemistry student can read in words. The candidate who can translate 'grams per litre times litres per minute, minus grams per litre times litres per minute' into dA/dt = c_in r_in - (A/V) r_out is doing the work the rubric is testing. The solution is then a separable equation, and the candidate should expect an exponential decay toward a steady state where the inflow concentration equals the outflow concentration. The steady state is A_∞ = c_in V for the constant-volume case, and the time to half the gap from the initial value to the steady state is (V / r_out) ln 2. Markers will usually award credit for a correct steady state even if the full solution is not obtained.

Worked micro-example: brine tank

Consider a 100-litre tank initially containing 50 grams of salt dissolved in 100 litres of water. Brine with 1 gram per litre enters at 3 litres per minute, the mixture is well stirred, and the mixture leaves at 3 litres per minute. The volume is constant at 100 litres, the inflow concentration is 1 gram per litre, the outflow concentration is A/100 grams per litre, and the differential equation is dA/dt = 3(1) - 3(A/100) = 3 - 0.03 A. The solution is the linear first-order form dA/dt + 0.03 A = 3, and the integrating factor is e^{0.03 t}. The general solution is A(t) = 100 + (A_0 - 100) e^{-0.03 t} = 100 - 50 e^{-0.03 t} grams. The candidate who reports the steady state of 100 grams is correct and earns the first two rubric points even if the integration is not completed. The candidate who reports the time to reach, say, 90 grams can solve 90 = 100 - 50 e^{-0.03 t} for t, which gives t = (1/0.03) ln 5, approximately 53.6 minutes. This answer is a one-line calculation once the differential equation is on the page.

Modelling, interpretation, and the marks that depend on the stem

The last section of the unit, and the part of the question that often decides between a 4 and a 5, is the interpretation of the solution. The exam asks the candidate to read the constant of integration against the initial condition, to find the time at which the population reaches a particular value, and to describe the long-term behaviour in words. The wording of the interpretation is marked strictly, and a candidate who writes 'the answer is 800' when the rubric expects 'the population is 800 grams at t = 0' loses credit. The marker is reading for three things: the units of the answer, the time at which the value is reached if one is asked for, and the qualitative behaviour of the solution as t grows without bound. Each is a separate point in the rubric, and each is recoverable independently.

For A-Level candidates using AP-style preparation, the overlap with the A-Level unit on differential equations is partial but useful. A-Level questions often ask for a general solution plus a particular constant, while AP questions often ask for the constant plus an interpretation. The candidate who can read a constant in context and write down its meaning in one sentence is doing the work that the rubric scores, and the preparation strategy that produces that habit is to write the interpretation on every practice problem, even when the rubric does not require it. In my experience this is the single most reliable way to convert a 4 into a 5.

Three reading habits that earn the interpretation points

The first habit is to write the units next to the answer. 'Eight hundred' is not '800 grams' or '800 bacteria', and the rubric is strict on this. The second habit is to write the time of the answer if the problem contains a t. The candidate who writes 'the population is 800 at t = 0' has read the question, while the candidate who writes 'the population is 800' has only done the algebra. The third habit is to use the language of the original stem. If the stem says 'a tank of brine', the answer should refer to the amount of salt in the tank, not to a generic 'y' or 'A'. These three habits are simple, they cost nothing in time, and they add a point on the average rubric.

Question types, scoring weights, and the allocation of study time

The differential-equations topic accounts for roughly 6 to 12 per cent of the AP Calculus AB exam and 6 to 9 per cent of the BC exam, distributed across multiple-choice and free-response items. The free-response items on the topic usually carry 6 to 9 raw points, distributed across two or three parts, and the topic is also embedded in questions that test the Fundamental Theorem of Calculus, the definite integral, and the average value of a function. The scoring logic is the same as the rest of the exam: each rubric point is awarded for a specific action, and the candidate's preparation strategy should map each practice problem to the rubric before the candidate ever picks up a pencil. The 4-to-5 grade band requires roughly 65 to 80 per cent of the available points, and a candidate who can score cleanly on the differential-equation items has a meaningful buffer for the harder BC topics.

Study time should be allocated with the question types in mind. Separable equations, slope fields, and Euler's method together account for the majority of the marks, and a candidate who is solid on these three families can earn a full score on the unit even if the integrating factor and the logistic integration are shaky. The integrating factor and the logistic are the topics that distinguish a 4 from a 5, and a candidate aiming at the 5 should spend half of the available study time on those two families. Mixture problems are a high-yield item for a candidate who has time, because the differential equation is essentially given in words and the integration is straightforward once the equation is on the page.

Suggested allocation across an A-Level-aligned prep cycle

For a 12-week preparation block that overlaps with A-Level work, week one and two should cover separable equations and exponential families, with one practice set per day and a full rubric review at the end of each week. Weeks three and four should cover slope fields and Euler's method, with two practice sets per week and an emphasis on the marker-awarded points. Weeks five and six should cover integrating factors and linear first-order equations, with one practice set per day for the integrating factor and a separate set for the linear case. Weeks seven and eight should cover logistic and mixture problems, with the same density of practice. Weeks nine and ten should consolidate with mixed-topic practice sets drawn from past exam papers, and weeks eleven and twelve should be timed full-paper practice with rubric review. The cycle is dense but it leaves room for the A-Level work that runs in parallel, and the differential-equations block is short enough that the candidate does not lose momentum on the rest of the syllabus.

Common pitfalls and how to avoid them

The mistakes that recur across sitting after sitting fall into a small set, and the preparation strategy that addresses them is to write each one down at the top of a practice paper and tick the list as the candidate works. The first is the missed constant of integration: a candidate writes y = e^{x²} instead of y = C e^{x²}, and the marker withholds a point. The second is the domain error: a candidate divides by y and forgets that y = 0 is a singular solution, and the marker withholds a point on the explicit form. The third is the wrong sign in the integrating factor: a candidate writes e^{-2x} when the standard form calls for e^{2x}, and the subsequent steps are not recoverable. The fourth is the missing units: a candidate writes 'the answer is 800' and the marker withholds a point on the interpretation. The fifth is the sign in the logistic solution: a candidate writes P(t) = M / (1 + A e^{kt}) instead of M / (1 + A e^{-kt}), and the long-term behaviour is wrong. The sixth is the unverified Euler's method: a candidate writes down the next point in the table but does not check the sign of the slope at the previous point, and the iteration drifts. The seventh is the verbal stem: a candidate answers the wrong question because they did not read the units, and the marker withholds all of the interpretation points.

The tactical response to each is the same: slow down at the rubric, write the first line, write the units, write the constant of integration, and check the long-term behaviour against the original differential equation. In my experience the candidate who adopts these five habits will lift their differential-equation score by 1.5 to 2 raw points per sitting, which is the difference between a 4 and a 5 in roughly 70 per cent of cases. The habits are not glamorous, and they will not feel like the heart of the calculus, but they are the habits that the marker rewards. Most candidates reading this who have lost marks on the topic have lost them on at least two of the seven pitfalls, and the fastest correction is to record the list on the inside cover of the practice book and to check it before the final answer is written down.

Pulling it together: how the topic is examined and how to read the paper

The differential-equations topic on the AP Calculus exam is a test of three things, in this order: the candidate's ability to translate a verbal description into a differential equation, the candidate's ability to solve that equation with a small menu of techniques, and the candidate's ability to read the solution back into the original context. The exam reuses the same six or seven families across sitting after sitting, and the candidate who can pattern-match a stem in five seconds has bought back a substantial amount of time for the more open-ended BC items. The scoring on the topic is generous in the sense that rubric points are partitioned so that a candidate can earn partial credit on almost any question, but strict in the sense that each point is awarded for a specific action and a missed action costs the point regardless of the surrounding work. Preparation strategy should respect that structure: practice to a rubric, write the units, write the constant of integration, and check the long-term behaviour.

For an A-Level-aligned student building a preparation plan that overlaps with AP-style work, the right starting point is a diagnostic on the four families covered above: separable, exponential, linear with integrating factor, and logistic. The diagnostic should be timed at 25 to 30 minutes, scored against the rubric of a real exam, and reviewed item by item. The candidate who can score 80 per cent on that diagnostic is in good shape for the 5 grade band on AP and the A* grade band at A-Level; the candidate who cannot should plan an additional four to six weeks of targeted practice on the failing families. The unit rewards targeted practice more than any other topic in the syllabus, because the families are narrow, the techniques are stable, and the rubric is predictable. A candidate who invests the time will see the return on the exam day.

TestPrep Europe's diagnostic on the differential-equations unit is a natural starting point for candidates who want to map their preparation to the rubric before the first timed paper.

Frequently asked questions