Concavity is the single topic on the AP Calculus exam where students write the most words, justify the least, and still lose a rubric row. The College Board asks about it in nearly every administration, both as a multiple-choice stem on the derivative interpretation side and as a free-response row that demands a written interval, a written justification, and a written identification of any inflection points. Most students who land a 4 instead of a 5 on AP Calculus AB or BC lose the points here for one of three reasons: they confuse concavity with monotonicity, they trust the first derivative where the second derivative was the right tool, or they hand the reader an interval with no sign chart to back it up. This article walks through the precise meaning of concave up and concave down, the way the AP Calculus rubric scores a concavity claim, and the preparation habits that turn a tentative answer into a defended one.
What concavity actually means on an AP Calculus answer
Concavity is a property of the graph, not the function. A function is concave up on an interval if its graph lies above every tangent line drawn at any point in that interval, and concave down if its graph lies below every tangent line. AP Calculus students often slip into saying the function is 'increasing' when the rubric wants 'concave up', and the two ideas are not interchangeable. A function can be decreasing and concave up, increasing and concave down, constant and linear (which is neither), or any combination the curve allows. On a free-response answer the rubric reads the words literally, so the distinction between 'the function is increasing on (a, b)' and 'the function is concave up on (a, b)' is the difference between a 1 and a 0 on the first derivative interpretation row.
The cleanest working definition for exam day is the tangent-line definition above, paired with the slope-of-tangent-line interpretation: concave up intervals are exactly those on which the tangent slopes are increasing, and concave down intervals are exactly those on which the tangent slopes are decreasing. That second view is the one that connects to differentiation, because 'tangent slopes are increasing' is a way of saying the first derivative is itself an increasing function, which on a differentiable curve is equivalent to the second derivative being positive. Most candidates reading this for the first time find the second view easier to defend in writing, because the rubric likes statements that end with a sign on f″.
Three landmarks anchor the topic for the exam. First, an inflection point is a point on the graph where the concavity actually changes, not just a point where f″(x) happens to equal zero. f″(c) = 0 with no sign change is a candidate only, never an inflection point. Second, a point of inflection must be a point on the curve, so the x-coordinate alone is not enough; the rubric wants (c, f(c)) or at least the x with the matching y. Third, concavity is a local-to-global hybrid: the function can be concave up on a closed interval, open interval, or even on a union, and any interval declared in the answer must be backed by the sign of f″ on that exact set.
AP Calculus BC candidates should remember that the same rules govern the analysis of parametric, vector, and polar curves, with derivatives taken in the appropriate form. The rubric on a concavity row of a BC-only prompt will still demand a written interval and a written justification; the calculus is just one extra step of differentiation away.
The second derivative test: when f″ decides a multiple-choice stem
On the multiple-choice section the College Board uses concavity mainly to test the second derivative test for local extrema. The test is short to state and easy to apply, but its conditions are exactly where careless candidates lose the point. The second derivative test says that if f′(c) = 0 and f″(c) is positive, then f has a local minimum at c; if f″(c) is negative, then f has a local maximum; and if f″(c) = 0, the test is inconclusive. The last clause is the one that costs the most points, because an inconclusive test is not a 'no' answer, it is an invitation to use the first derivative test or the function values directly.
For most candidates, the second derivative test is the fastest way to classify a critical point on a multiple-choice stem, provided the second derivative is easy to compute at the critical point. In my experience, students overuse the test on cubic, quartic, and rational functions where the second derivative is simple, then suddenly switch to it on a stem with ln x or e^(kx) where the first derivative test would have been two lines faster. The honest decision rule: pick the test that finishes inside 90 seconds, and that is almost always the second derivative test on polynomial stems and the first derivative test on transcendental stems.
Two further traps appear on the MCQ. First, a stem may give f′ and ask about f″, and the candidate is supposed to differentiate, not look at the sign of f′. Reading the prompt carefully is half the work on a 2-minute stem. Second, the test is inconclusive only when f″(c) = 0, not when f″(c) is undefined. A vertical tangent with f″ undefined can still host a clean local extremum, and the rubric on a free-response will accept a sign chart on f′ as the defence.
Building a sign chart that earns the AP Calculus rubric point
The single highest-leverage habit for a concavity row on an FRQ is to draw a sign chart of f″. The rubric is unforgiving: a written interval with no supporting sign of f″ on that interval is, in my experience, treated as an unsupported claim and earns at most a partial point for identifying the interval but not the justification. The sign chart itself is fast: list the critical values of f″, the points where f″ is undefined, and any x-values the prompt supplies, then test one number in each open interval to see whether f″ is positive or negative there.
Consider a worked example. Suppose f(x) = x^4 − 4x^3. Then f′(x) = 4x^3 − 12x^2, f″(x) = 12x^2 − 24x = 12x(x − 2). The zeros of f″ are x = 0 and x = 2, and f″ is a polynomial so it is defined everywhere. Test x = −1: f″(−1) = 12(−1)(−3) = 36, positive. Test x = 1: f″(1) = 12(1)(−1) = −12, negative. Test x = 3: f″(3) = 12(3)(1) = 36, positive. The sign chart is therefore +, −, +, and the function is concave up on (−∞, 0), concave down on (0, 2), and concave up on (2, ∞). The points (0, f(0)) = (0, 0) and (2, f(2)) = (2, −16) are the inflection points, because the concavity actually changes sign at each.
Notice the small moves that add up on the rubric. The chart must include the actual y-value at the candidate inflection x, not just the x. The interval endpoints should be open or closed to match the type of the critical value, and if the prompt supplied a domain, the interval must respect it. The justification in the answer box usually says something like 'f is concave up on (0, 2) because f″(x) = 12x(x − 2) < 0 on (0, 2)'. That sentence pattern — interval, then derivative with a sign claim — is what the rubric language rewards.
Concavity FRQ prompt shapes you should recognise on sight
There are four FRQ prompt shapes that test concavity on AB and BC, and each one wants a slightly different answer.
1. The direct claim. The prompt says 'Find the intervals on which f is concave up and concave down' and the answer wants intervals plus a written sign on f″ for each. This is the cleanest shape and the one that most preparation books lead with.
2. The implicit-curve prompt. The prompt gives an equation in x and y such as x^2 + y^2 = 1 or xy = 4 and asks about concavity of y as a function of x on a given interval. The answer needs dy/dx and d^2y/dx^2 from implicit differentiation, then a sign chart on d^2y/dx^2 restricted to the supplied x-interval. AP Calculus BC candidates see this shape with a parametric pair (x(t), y(t)) and the second derivative formula d^2y/dx^2 = (d/dt (dy/dx)) / (dx/dt).
3. The interpretation row. The prompt describes a context, such as a particle's position or a population model, and asks whether the rate of change of the rate of change is positive or negative on a window. The answer wants a sign on f″, framed in the language of the context: 'the velocity is increasing on the interval because the acceleration is positive'.
4. The function-table row. The prompt supplies a partially labelled table of f, f′, f″ and asks for an inflection interval or a concavity claim. The candidate is supposed to read the table, not recompute derivatives. This shape rewards students who look at where the sign of f″ flips, and it is one of the easiest rows on the exam for a student who has practised sign-chart reading.
Common pitfalls on these four shapes include the following, all of which I have watched cost a rubric row in past administrations:
- Writing 'concave up at x = 0' instead of 'concave up on (a, b)'. Concavity is an interval property, not a point property, and the rubric wants an interval.
- Listing the critical values of f″ as inflection points without checking that the sign actually changes. f″ = 0 is a candidate, not a guarantee.
- Forgetting the y-coordinate at an inflection point. The rubric on AB treats a missing y as a partial point; on BC the same row often shares its justification with a derivative, so the missing y can collapse two points into one.
- Confusing the sign of f′ with the sign of f″. A positive f′ says the function is increasing; a positive f″ says it is concave up. They are different rows of the rubric and a candidate who swaps them usually scores 0 on both.
- Closing an interval at a vertical tangent where f″ is undefined, when the domain actually extends through that point. The interval is open at the undefined point, but the function is still defined there, and the rubric wants the cleanest defensible interval.
Concavity on a particle-motion FRQ: the velocity-acceleration bridge
Particle-motion prompts are the second most common place where concavity appears, and the bridge between calculus and physics is short. If s(t) is position, v(t) = s′(t) is velocity, and a(t) = v′(t) = s″(t) is acceleration. The function s is concave up exactly when a(t) is positive, regardless of the sign of v(t). The interpretation on a typical AP Calculus FRQ is: 'the particle is moving to the right and speeding up on the interval' or 'the particle is moving to the left and slowing down', and the candidate is expected to combine the sign of v with the sign of a to reach the right physical description.
A clean worked example. Let s(t) = t^3 − 6t^2 on the interval [0, 5]. Then v(t) = 3t^2 − 12t = 3t(t − 4), and a(t) = 6t − 12. Setting v = 0 gives t = 0 and t = 4; setting a = 0 gives t = 2. A sign chart on v gives the motion: positive for t < 0, negative on (0, 4), positive on (4, 5); on the visible domain the particle is moving left on (0, 4) and right on (4, 5). A sign chart on a gives the concavity: a(t) < 0 on (0, 2) and a(t) > 0 on (2, 5). Putting the two together, on (0, 2) the particle is moving left and slowing down (negative velocity, negative acceleration, so speed is decreasing — actually increasing, so the candidate should say 'speeding up' since both v and a are negative). On (2, 4) the particle is moving left and speeding up, since v < 0 and a > 0 implies speed is increasing. The point t = 2 is the inflection point of s, and the row the rubric usually asks is 'find the t at which the particle changes from speeding up to slowing down', which is the inflection of s, not the critical point of s.
The most common error on this style of question is to treat the critical point of s as the answer. The critical point of s is where v = 0, which is where the particle reverses direction, not where it changes speed regime. The change in speed regime is at the inflection of s, which is where a = 0 with a sign change. The rubric distinguishes these two rows, and a candidate who conflates them loses a row.
Reading an MCQ concavity stem in under 90 seconds
The MCQ section rewards a specific kind of fluency: reading the stem, choosing the right test, and defending the answer in the head before scanning the choices. The best preparation for the concavity stems on the MCQ is to drill the second derivative test on simple polynomials until it becomes a reflex, then transfer that reflex to the four shapes below. None of these shapes should take more than 90 seconds on exam day.
Shape A: a stem gives f(x) and asks where the graph is concave up. The candidate differentiates twice, finds the zeros of f″, builds a quick sign chart, and matches the intervals to the choices. The trap is to read the sign of f′ instead of f″; on a 2-minute stem, that mistake costs the question and bleeds into the next one.
Shape B: a stem gives a graph and asks which interval is concave down. The candidate looks for tangent lines that sit above the curve on the interval, not for the function's monotonicity. A 'downward' slope on the graph is monotonicity, not concavity, and the rubric considers the two ideas different.
Shape C: a stem gives a table of f, f′, f″ and asks for the location of an inflection point. The candidate scans the row for f″ and looks for the place where the sign of f″ flips, then matches the x to the table. The trap is to pick a zero of f″ without checking the sign change, which is exactly the kind of half-credit bait the College Board likes to seed.
Shape D: a stem gives f′ and asks for concavity of f. The candidate differentiates f′ to get f″, then proceeds as in shape A. The trap is to treat the zeros of f′ as the answer, which gives critical points of f, not concavity intervals.
| Stem shape | First step | Final check |
|---|---|---|
| Shape A: f(x) given, concavity asked | Differentiate twice; find zeros of f″ | Sign chart of f″ across each interval |
| Shape B: graph given, concavity asked | Inspect tangent-line position, not slope | Concave up intervals are above tangents |
| Shape C: table of f, f′, f″ given | Locate sign change in f″ row | Match x to the candidate inflection value |
| Shape D: f′ given, concavity of f asked | Compute f″ from f′ | Apply sign chart to f″ |
AP Calculus BC only: concavity on parametric and polar curves
BC candidates see concavity appear on FRQs where x and y are given parametrically, in vector form, or in polar form. The pattern is identical to the AB rules: the answer wants intervals where d^2y/dx^2 is positive or negative, and a sign chart on d^2y/dx^2 is the supporting evidence. The complication is the formula. For a parametric pair (x(t), y(t)) with dx/dt ≠ 0, the first derivative is dy/dx = (dy/dt) / (dx/dt), and the second derivative is d^2y/dx^2 = (d/dt (dy/dx)) / (dx/dt). On an FRQ the candidate is expected to write the second derivative, simplify, and then build a sign chart over the supplied t-interval, not the x-interval.
A worked BC example. Let x(t) = t^2 − 1 and y(t) = t^3 − t for t in [−2, 2]. Then dx/dt = 2t, dy/dt = 3t^2 − 1, so dy/dx = (3t^2 − 1) / (2t). Differentiate with respect to t: d/dt (dy/dx) = (6t · 2t − (3t^2 − 1) · 2) / (4t^2) = (12t^2 − 6t^2 + 2) / (4t^2) = (6t^2 + 2) / (4t^2). Then d^2y/dx^2 = (6t^2 + 2) / (4t^2) · 1/(2t) = (6t^2 + 2) / (8t^3). The sign of d^2y/dx^2 is the sign of (6t^2 + 2) / t^3. The numerator is always positive, so the sign is the sign of 1/t^3, which is positive for t > 0 and negative for t < 0. The curve is concave down for t < 0, concave up for t > 0, and the curve has an inflection at t = 0 (where dx/dt = 0, so the formula fails, and the candidate should say so explicitly).
For polar curves the BC rubric still expects a sign on d^2y/dx^2 derived from the polar conversion. For r = f(θ), the candidate writes x = r cos θ, y = r sin θ, computes dy/dθ, dx/dθ, and proceeds as in the parametric case. The dx/dθ ≠ 0 caveat applies on every interval where the curve passes through the origin with a vertical tangent, and the rubric wants the candidate to call out the failure rather than compute through it.
Preparation strategy: how to drill concavity in the final four weeks
The most efficient way to prepare for the concavity rows is a four-week plan that pairs a sign-chart drill with a rubric-language drill. In week one, the student does 10 to 12 AB-level FRQ rows where the prompt gives a function and asks for intervals of concavity. The student writes the answer in the form 'concave up on (a, b) because f″(x) = … is positive on (a, b)', then checks the answer against the official scoring guidelines. In week two, the student does 8 to 10 BC-level rows on parametric or polar curves, with the same rubric-language drill. In week three, the student tackles 15 to 20 multiple-choice stems in the four shapes above, timed at 90 seconds each. In week four, the student takes a full mock exam and grades the concavity rows separately, then returns to the rubric for any row that scored less than full.
Scoring-wise, concavity is not a separate question on the AP Calculus exam; it is a row inside a longer FRQ or a stem inside a derivative interpretation MCQ cluster. The cumulative weight is significant, and most students who land a 4 instead of a 5 lose the points across multiple rows of the same exam, not in one catastrophic question. A student who can defend a concavity claim with a sign chart, write the interval with the right endpoint type, and identify an inflection point with its y-coordinate is in a strong position to convert a 4 to a 5. The exam format itself is 45 multiple-choice questions and 6 free-response questions over 3 hours and 15 minutes, and a paced candidate will have spent roughly 25 minutes total on the rows that test concavity across the FRQ section, plus a handful of MCQ stems in the derivative interpretation cluster. That is a real slice of the score, and a focused preparation plan on the topic will move the needle more than an unfocused re-read of the whole course.
Conclusion and next steps
Concavity is the part of AP Calculus where the rubric is most literal and the candidate's writing is most exposed. The path to a defended answer is short: build a sign chart of f″, write the interval, name the inflection point with its y-coordinate, and use the rubric language 'concave up on (a, b) because f″(x) > 0 on (a, b)'. BC candidates extend the same pattern to parametric and polar curves by carrying the second derivative formula and respecting the dx/dt ≠ 0 caveat. A four-week plan that pairs 10 to 12 AB rows, 8 to 10 BC rows, 15 to 20 timed MCQ stems, and a graded mock exam is the shortest path to a defended answer on exam day. AP Courses' one-to-one AP Calculus BC programme analyses each student's FRQ concavity rows against the official scoring guidelines and turns a 5 target into a sign-chart drill plan keyed to the student's actual error pattern.